From 8b6cf128af65db9faca8529f680c0f24f9660377 Mon Sep 17 00:00:00 2001
From: Martin Holst Swende <martin@swende.se>
Date: Wed, 13 Sep 2023 19:13:47 +0200
Subject: [PATCH] metrics: refactor metrics (#28035)

This change includes a lot of things, listed below.

### Split up interfaces, write vs read

The interfaces have been split up into one write-interface and one read-interface, with `Snapshot` being the gateway from write to read. This simplifies the semantics _a lot_.

Example of splitting up an interface into one readonly 'snapshot' part, and one updatable writeonly part:

```golang
type MeterSnapshot interface {
	Count() int64
	Rate1() float64
	Rate5() float64
	Rate15() float64
	RateMean() float64
}

// Meters count events to produce exponentially-weighted moving average rates
// at one-, five-, and fifteen-minutes and a mean rate.
type Meter interface {
	Mark(int64)
	Snapshot() MeterSnapshot
	Stop()
}
```

### A note about concurrency

This PR makes the concurrency model clearer. We have actual meters and snapshot of meters. The `meter` is the thing which can be accessed from the registry, and updates can be made to it.

- For all `meters`, (`Gauge`, `Timer` etc), it is assumed that they are accessed by different threads, making updates. Therefore, all `meters` update-methods (`Inc`, `Add`, `Update`, `Clear` etc) need to be concurrency-safe.
- All `meters` have a `Snapshot()` method. This method is _usually_ called from one thread, a backend-exporter. But it's fully possible to have several exporters simultaneously: therefore this method should also be concurrency-safe.

TLDR: `meter`s are accessible via registry, all their methods must be concurrency-safe.

For all `Snapshot`s, it is assumed that an individual exporter-thread has obtained a `meter` from the registry, and called the `Snapshot` method to obtain a readonly snapshot. This snapshot is _not_ guaranteed to be concurrency-safe. There's no need for a snapshot to be concurrency-safe, since exporters should not share snapshots.

Note, though: that by happenstance a lot of the snapshots _are_ concurrency-safe, being unmutable minimal representations of a value. Only the more complex ones are _not_ threadsafe, those that lazily calculate things like `Variance()`, `Mean()`.

Example of how a background exporter typically works, obtaining the snapshot and sequentially accessing the non-threadsafe methods in it:
```golang
		ms := metric.Snapshot()
                ...
		fields := map[string]interface{}{
			"count":    ms.Count(),
			"max":      ms.Max(),
			"mean":     ms.Mean(),
			"min":      ms.Min(),
			"stddev":   ms.StdDev(),
			"variance": ms.Variance(),
```

TLDR: `snapshots` are not guaranteed to be concurrency-safe (but often are).

### Sample changes

I also changed the `Sample` type: previously, it iterated the samples fully every time `Mean()`,`Sum()`, `Min()` or `Max()` was invoked. Since we now have readonly base data, we can just iterate it once, in the constructor, and set all four values at once.

The same thing has been done for runtimehistogram.

### ResettingTimer API

Back when ResettingTImer was implemented, as part of https://github.com/ethereum/go-ethereum/pull/15910, Anton implemented a `Percentiles` on the new type. However, the method did not conform to the other existing types which also had a `Percentiles`.

1. The existing ones, on input, took `0.5` to mean `50%`. Anton used `50` to mean `50%`.
2. The existing ones returned `float64` outputs, thus interpolating between values. A value-set of `0, 10`, at `50%` would return `5`, whereas Anton's would return either `0` or `10`.

This PR removes the 'new' version, and uses only the 'legacy' percentiles, also for the ResettingTimer type.

The resetting timer snapshot was also defined so that it would expose the internal values. This has been removed, and getters for `Max, Min, Mean` have been added instead.

### Unexport types

A lot of types were exported, but do not need to be. This PR unexports quite a lot of them.
---
 core/state/statedb.go                       |  10 +-
 metrics/counter.go                          |  72 +---
 metrics/counter_float64.go                  |  61 +--
 metrics/counter_float_64_test.go            |  16 +-
 metrics/counter_test.go                     |  14 +-
 metrics/doc.go                              |   4 -
 metrics/ewma.go                             |  91 ++---
 metrics/ewma_test.go                        | 233 +++--------
 metrics/exp/exp.go                          |  32 +-
 metrics/gauge.go                            | 116 ++----
 metrics/gauge_float64.go                    |  80 +---
 metrics/gauge_float64_test.go               |  33 +-
 metrics/gauge_info.go                       |  88 +---
 metrics/gauge_info_test.go                  |  49 +--
 metrics/gauge_test.go                       |  39 +-
 metrics/graphite.go                         |  10 +-
 metrics/histogram.go                        | 158 +------
 metrics/histogram_test.go                   |   8 +-
 metrics/inactive.go                         |  48 +++
 metrics/influxdb/influxdb.go                |  21 +-
 metrics/influxdb/influxdb_test.go           |   2 +-
 metrics/influxdb/testdata/influxdbv1.want   |   4 +-
 metrics/influxdb/testdata/influxdbv2.want   |   4 +-
 metrics/internal/sampledata.go              |  33 +-
 metrics/internal/sampledata_test.go         |  27 ++
 metrics/librato/librato.go                  |  73 ++--
 metrics/log.go                              |  10 +-
 metrics/meter.go                            | 207 ++--------
 metrics/meter_test.go                       |  30 +-
 metrics/metrics.go                          |   6 +
 metrics/metrics_test.go                     |   4 +-
 metrics/opentsdb.go                         |  10 +-
 metrics/opentsdb_test.go                    |  15 +
 metrics/prometheus/collector.go             |  25 +-
 metrics/prometheus/collector_test.go        |   4 +-
 metrics/prometheus/testdata/prometheus.want |  28 +-
 metrics/registry.go                         |   8 +-
 metrics/registry_test.go                    |   4 +-
 metrics/resetting_sample.go                 |   2 +-
 metrics/resetting_timer.go                  | 204 ++++-----
 metrics/resetting_timer_test.go             | 101 ++---
 metrics/runtimehistogram.go                 | 218 +++++-----
 metrics/runtimehistogram_test.go            |  37 +-
 metrics/sample.go                           | 432 ++++++--------------
 metrics/sample_test.go                      | 150 +++----
 metrics/syslog.go                           |  10 +-
 metrics/testdata/opentsb.want               |   2 +-
 metrics/timer.go                            | 214 ++--------
 metrics/timer_test.go                       |  10 +-
 metrics/writer.go                           |  10 +-
 50 files changed, 1035 insertions(+), 2032 deletions(-)
 delete mode 100644 metrics/doc.go
 create mode 100644 metrics/inactive.go
 create mode 100644 metrics/internal/sampledata_test.go

diff --git a/core/state/statedb.go b/core/state/statedb.go
index bd578ba23d..c1b5b0874c 100644
--- a/core/state/statedb.go
+++ b/core/state/statedb.go
@@ -1061,12 +1061,10 @@ func (s *StateDB) deleteStorage(addr common.Address, addrHash common.Hash, root
 			slotDeletionSkip.Inc(1)
 		}
 		n := int64(len(slots))
-		if n > slotDeletionMaxCount.Value() {
-			slotDeletionMaxCount.Update(n)
-		}
-		if int64(size) > slotDeletionMaxSize.Value() {
-			slotDeletionMaxSize.Update(int64(size))
-		}
+
+		slotDeletionMaxCount.UpdateIfGt(int64(len(slots)))
+		slotDeletionMaxSize.UpdateIfGt(int64(size))
+
 		slotDeletionTimer.UpdateSince(start)
 		slotDeletionCount.Mark(n)
 		slotDeletionSize.Mark(int64(size))
diff --git a/metrics/counter.go b/metrics/counter.go
index 55e1c59540..cb81599c21 100644
--- a/metrics/counter.go
+++ b/metrics/counter.go
@@ -4,13 +4,16 @@ import (
 	"sync/atomic"
 )
 
+type CounterSnapshot interface {
+	Count() int64
+}
+
 // Counters hold an int64 value that can be incremented and decremented.
 type Counter interface {
 	Clear()
-	Count() int64
 	Dec(int64)
 	Inc(int64)
-	Snapshot() Counter
+	Snapshot() CounterSnapshot
 }
 
 // GetOrRegisterCounter returns an existing Counter or constructs and registers
@@ -38,13 +41,13 @@ func NewCounter() Counter {
 	if !Enabled {
 		return NilCounter{}
 	}
-	return &StandardCounter{}
+	return new(StandardCounter)
 }
 
 // NewCounterForced constructs a new StandardCounter and returns it no matter if
 // the global switch is enabled or not.
 func NewCounterForced() Counter {
-	return &StandardCounter{}
+	return new(StandardCounter)
 }
 
 // NewRegisteredCounter constructs and registers a new StandardCounter.
@@ -70,75 +73,40 @@ func NewRegisteredCounterForced(name string, r Registry) Counter {
 	return c
 }
 
-// CounterSnapshot is a read-only copy of another Counter.
-type CounterSnapshot int64
-
-// Clear panics.
-func (CounterSnapshot) Clear() {
-	panic("Clear called on a CounterSnapshot")
-}
+// counterSnapshot is a read-only copy of another Counter.
+type counterSnapshot int64
 
 // Count returns the count at the time the snapshot was taken.
-func (c CounterSnapshot) Count() int64 { return int64(c) }
-
-// Dec panics.
-func (CounterSnapshot) Dec(int64) {
-	panic("Dec called on a CounterSnapshot")
-}
-
-// Inc panics.
-func (CounterSnapshot) Inc(int64) {
-	panic("Inc called on a CounterSnapshot")
-}
-
-// Snapshot returns the snapshot.
-func (c CounterSnapshot) Snapshot() Counter { return c }
+func (c counterSnapshot) Count() int64 { return int64(c) }
 
 // NilCounter is a no-op Counter.
 type NilCounter struct{}
 
-// Clear is a no-op.
-func (NilCounter) Clear() {}
-
-// Count is a no-op.
-func (NilCounter) Count() int64 { return 0 }
-
-// Dec is a no-op.
-func (NilCounter) Dec(i int64) {}
-
-// Inc is a no-op.
-func (NilCounter) Inc(i int64) {}
-
-// Snapshot is a no-op.
-func (NilCounter) Snapshot() Counter { return NilCounter{} }
+func (NilCounter) Clear()                    {}
+func (NilCounter) Dec(i int64)               {}
+func (NilCounter) Inc(i int64)               {}
+func (NilCounter) Snapshot() CounterSnapshot { return (*emptySnapshot)(nil) }
 
 // StandardCounter is the standard implementation of a Counter and uses the
 // sync/atomic package to manage a single int64 value.
-type StandardCounter struct {
-	count atomic.Int64
-}
+type StandardCounter atomic.Int64
 
 // Clear sets the counter to zero.
 func (c *StandardCounter) Clear() {
-	c.count.Store(0)
-}
-
-// Count returns the current count.
-func (c *StandardCounter) Count() int64 {
-	return c.count.Load()
+	(*atomic.Int64)(c).Store(0)
 }
 
 // Dec decrements the counter by the given amount.
 func (c *StandardCounter) Dec(i int64) {
-	c.count.Add(-i)
+	(*atomic.Int64)(c).Add(-i)
 }
 
 // Inc increments the counter by the given amount.
 func (c *StandardCounter) Inc(i int64) {
-	c.count.Add(i)
+	(*atomic.Int64)(c).Add(i)
 }
 
 // Snapshot returns a read-only copy of the counter.
-func (c *StandardCounter) Snapshot() Counter {
-	return CounterSnapshot(c.Count())
+func (c *StandardCounter) Snapshot() CounterSnapshot {
+	return counterSnapshot((*atomic.Int64)(c).Load())
 }
diff --git a/metrics/counter_float64.go b/metrics/counter_float64.go
index d1197bb8e0..15c81494ef 100644
--- a/metrics/counter_float64.go
+++ b/metrics/counter_float64.go
@@ -5,13 +5,16 @@ import (
 	"sync/atomic"
 )
 
+type CounterFloat64Snapshot interface {
+	Count() float64
+}
+
 // CounterFloat64 holds a float64 value that can be incremented and decremented.
 type CounterFloat64 interface {
 	Clear()
-	Count() float64
 	Dec(float64)
 	Inc(float64)
-	Snapshot() CounterFloat64
+	Snapshot() CounterFloat64Snapshot
 }
 
 // GetOrRegisterCounterFloat64 returns an existing CounterFloat64 or constructs and registers
@@ -71,47 +74,19 @@ func NewRegisteredCounterFloat64Forced(name string, r Registry) CounterFloat64 {
 	return c
 }
 
-// CounterFloat64Snapshot is a read-only copy of another CounterFloat64.
-type CounterFloat64Snapshot float64
-
-// Clear panics.
-func (CounterFloat64Snapshot) Clear() {
-	panic("Clear called on a CounterFloat64Snapshot")
-}
+// counterFloat64Snapshot is a read-only copy of another CounterFloat64.
+type counterFloat64Snapshot float64
 
 // Count returns the value at the time the snapshot was taken.
-func (c CounterFloat64Snapshot) Count() float64 { return float64(c) }
-
-// Dec panics.
-func (CounterFloat64Snapshot) Dec(float64) {
-	panic("Dec called on a CounterFloat64Snapshot")
-}
+func (c counterFloat64Snapshot) Count() float64 { return float64(c) }
 
-// Inc panics.
-func (CounterFloat64Snapshot) Inc(float64) {
-	panic("Inc called on a CounterFloat64Snapshot")
-}
-
-// Snapshot returns the snapshot.
-func (c CounterFloat64Snapshot) Snapshot() CounterFloat64 { return c }
-
-// NilCounterFloat64 is a no-op CounterFloat64.
 type NilCounterFloat64 struct{}
 
-// Clear is a no-op.
-func (NilCounterFloat64) Clear() {}
-
-// Count is a no-op.
-func (NilCounterFloat64) Count() float64 { return 0.0 }
-
-// Dec is a no-op.
-func (NilCounterFloat64) Dec(i float64) {}
-
-// Inc is a no-op.
-func (NilCounterFloat64) Inc(i float64) {}
-
-// Snapshot is a no-op.
-func (NilCounterFloat64) Snapshot() CounterFloat64 { return NilCounterFloat64{} }
+func (NilCounterFloat64) Clear()                           {}
+func (NilCounterFloat64) Count() float64                   { return 0.0 }
+func (NilCounterFloat64) Dec(i float64)                    {}
+func (NilCounterFloat64) Inc(i float64)                    {}
+func (NilCounterFloat64) Snapshot() CounterFloat64Snapshot { return NilCounterFloat64{} }
 
 // StandardCounterFloat64 is the standard implementation of a CounterFloat64 and uses the
 // atomic to manage a single float64 value.
@@ -124,11 +99,6 @@ func (c *StandardCounterFloat64) Clear() {
 	c.floatBits.Store(0)
 }
 
-// Count returns the current value.
-func (c *StandardCounterFloat64) Count() float64 {
-	return math.Float64frombits(c.floatBits.Load())
-}
-
 // Dec decrements the counter by the given amount.
 func (c *StandardCounterFloat64) Dec(v float64) {
 	atomicAddFloat(&c.floatBits, -v)
@@ -140,8 +110,9 @@ func (c *StandardCounterFloat64) Inc(v float64) {
 }
 
 // Snapshot returns a read-only copy of the counter.
-func (c *StandardCounterFloat64) Snapshot() CounterFloat64 {
-	return CounterFloat64Snapshot(c.Count())
+func (c *StandardCounterFloat64) Snapshot() CounterFloat64Snapshot {
+	v := math.Float64frombits(c.floatBits.Load())
+	return counterFloat64Snapshot(v)
 }
 
 func atomicAddFloat(fbits *atomic.Uint64, v float64) {
diff --git a/metrics/counter_float_64_test.go b/metrics/counter_float_64_test.go
index f17aca330c..c21bd3307f 100644
--- a/metrics/counter_float_64_test.go
+++ b/metrics/counter_float_64_test.go
@@ -27,7 +27,7 @@ func BenchmarkCounterFloat64Parallel(b *testing.B) {
 		}()
 	}
 	wg.Wait()
-	if have, want := c.Count(), 10.0*float64(b.N); have != want {
+	if have, want := c.Snapshot().Count(), 10.0*float64(b.N); have != want {
 		b.Fatalf("have %f want %f", have, want)
 	}
 }
@@ -36,7 +36,7 @@ func TestCounterFloat64Clear(t *testing.T) {
 	c := NewCounterFloat64()
 	c.Inc(1.0)
 	c.Clear()
-	if count := c.Count(); count != 0 {
+	if count := c.Snapshot().Count(); count != 0 {
 		t.Errorf("c.Count(): 0 != %v\n", count)
 	}
 }
@@ -44,7 +44,7 @@ func TestCounterFloat64Clear(t *testing.T) {
 func TestCounterFloat64Dec1(t *testing.T) {
 	c := NewCounterFloat64()
 	c.Dec(1.0)
-	if count := c.Count(); count != -1.0 {
+	if count := c.Snapshot().Count(); count != -1.0 {
 		t.Errorf("c.Count(): -1.0 != %v\n", count)
 	}
 }
@@ -52,7 +52,7 @@ func TestCounterFloat64Dec1(t *testing.T) {
 func TestCounterFloat64Dec2(t *testing.T) {
 	c := NewCounterFloat64()
 	c.Dec(2.0)
-	if count := c.Count(); count != -2.0 {
+	if count := c.Snapshot().Count(); count != -2.0 {
 		t.Errorf("c.Count(): -2.0 != %v\n", count)
 	}
 }
@@ -60,7 +60,7 @@ func TestCounterFloat64Dec2(t *testing.T) {
 func TestCounterFloat64Inc1(t *testing.T) {
 	c := NewCounterFloat64()
 	c.Inc(1.0)
-	if count := c.Count(); count != 1.0 {
+	if count := c.Snapshot().Count(); count != 1.0 {
 		t.Errorf("c.Count(): 1.0 != %v\n", count)
 	}
 }
@@ -68,7 +68,7 @@ func TestCounterFloat64Inc1(t *testing.T) {
 func TestCounterFloat64Inc2(t *testing.T) {
 	c := NewCounterFloat64()
 	c.Inc(2.0)
-	if count := c.Count(); count != 2.0 {
+	if count := c.Snapshot().Count(); count != 2.0 {
 		t.Errorf("c.Count(): 2.0 != %v\n", count)
 	}
 }
@@ -85,7 +85,7 @@ func TestCounterFloat64Snapshot(t *testing.T) {
 
 func TestCounterFloat64Zero(t *testing.T) {
 	c := NewCounterFloat64()
-	if count := c.Count(); count != 0 {
+	if count := c.Snapshot().Count(); count != 0 {
 		t.Errorf("c.Count(): 0 != %v\n", count)
 	}
 }
@@ -93,7 +93,7 @@ func TestCounterFloat64Zero(t *testing.T) {
 func TestGetOrRegisterCounterFloat64(t *testing.T) {
 	r := NewRegistry()
 	NewRegisteredCounterFloat64("foo", r).Inc(47.0)
-	if c := GetOrRegisterCounterFloat64("foo", r); c.Count() != 47.0 {
+	if c := GetOrRegisterCounterFloat64("foo", r).Snapshot(); c.Count() != 47.0 {
 		t.Fatal(c)
 	}
 }
diff --git a/metrics/counter_test.go b/metrics/counter_test.go
index af26ef1548..1b15b23f21 100644
--- a/metrics/counter_test.go
+++ b/metrics/counter_test.go
@@ -14,7 +14,7 @@ func TestCounterClear(t *testing.T) {
 	c := NewCounter()
 	c.Inc(1)
 	c.Clear()
-	if count := c.Count(); count != 0 {
+	if count := c.Snapshot().Count(); count != 0 {
 		t.Errorf("c.Count(): 0 != %v\n", count)
 	}
 }
@@ -22,7 +22,7 @@ func TestCounterClear(t *testing.T) {
 func TestCounterDec1(t *testing.T) {
 	c := NewCounter()
 	c.Dec(1)
-	if count := c.Count(); count != -1 {
+	if count := c.Snapshot().Count(); count != -1 {
 		t.Errorf("c.Count(): -1 != %v\n", count)
 	}
 }
@@ -30,7 +30,7 @@ func TestCounterDec1(t *testing.T) {
 func TestCounterDec2(t *testing.T) {
 	c := NewCounter()
 	c.Dec(2)
-	if count := c.Count(); count != -2 {
+	if count := c.Snapshot().Count(); count != -2 {
 		t.Errorf("c.Count(): -2 != %v\n", count)
 	}
 }
@@ -38,7 +38,7 @@ func TestCounterDec2(t *testing.T) {
 func TestCounterInc1(t *testing.T) {
 	c := NewCounter()
 	c.Inc(1)
-	if count := c.Count(); count != 1 {
+	if count := c.Snapshot().Count(); count != 1 {
 		t.Errorf("c.Count(): 1 != %v\n", count)
 	}
 }
@@ -46,7 +46,7 @@ func TestCounterInc1(t *testing.T) {
 func TestCounterInc2(t *testing.T) {
 	c := NewCounter()
 	c.Inc(2)
-	if count := c.Count(); count != 2 {
+	if count := c.Snapshot().Count(); count != 2 {
 		t.Errorf("c.Count(): 2 != %v\n", count)
 	}
 }
@@ -63,7 +63,7 @@ func TestCounterSnapshot(t *testing.T) {
 
 func TestCounterZero(t *testing.T) {
 	c := NewCounter()
-	if count := c.Count(); count != 0 {
+	if count := c.Snapshot().Count(); count != 0 {
 		t.Errorf("c.Count(): 0 != %v\n", count)
 	}
 }
@@ -71,7 +71,7 @@ func TestCounterZero(t *testing.T) {
 func TestGetOrRegisterCounter(t *testing.T) {
 	r := NewRegistry()
 	NewRegisteredCounter("foo", r).Inc(47)
-	if c := GetOrRegisterCounter("foo", r); c.Count() != 47 {
+	if c := GetOrRegisterCounter("foo", r).Snapshot(); c.Count() != 47 {
 		t.Fatal(c)
 	}
 }
diff --git a/metrics/doc.go b/metrics/doc.go
deleted file mode 100644
index 13f429c168..0000000000
--- a/metrics/doc.go
+++ /dev/null
@@ -1,4 +0,0 @@
-package metrics
-
-const epsilon = 0.0000000000000001
-const epsilonPercentile = .00000000001
diff --git a/metrics/ewma.go b/metrics/ewma.go
index ed95cba19b..1d7a4f00cf 100644
--- a/metrics/ewma.go
+++ b/metrics/ewma.go
@@ -7,11 +7,14 @@ import (
 	"time"
 )
 
+type EWMASnapshot interface {
+	Rate() float64
+}
+
 // EWMAs continuously calculate an exponentially-weighted moving average
 // based on an outside source of clock ticks.
 type EWMA interface {
-	Rate() float64
-	Snapshot() EWMA
+	Snapshot() EWMASnapshot
 	Tick()
 	Update(int64)
 }
@@ -36,40 +39,19 @@ func NewEWMA15() EWMA {
 	return NewEWMA(1 - math.Exp(-5.0/60.0/15))
 }
 
-// EWMASnapshot is a read-only copy of another EWMA.
-type EWMASnapshot float64
+// ewmaSnapshot is a read-only copy of another EWMA.
+type ewmaSnapshot float64
 
 // Rate returns the rate of events per second at the time the snapshot was
 // taken.
-func (a EWMASnapshot) Rate() float64 { return float64(a) }
-
-// Snapshot returns the snapshot.
-func (a EWMASnapshot) Snapshot() EWMA { return a }
-
-// Tick panics.
-func (EWMASnapshot) Tick() {
-	panic("Tick called on an EWMASnapshot")
-}
-
-// Update panics.
-func (EWMASnapshot) Update(int64) {
-	panic("Update called on an EWMASnapshot")
-}
+func (a ewmaSnapshot) Rate() float64 { return float64(a) }
 
 // NilEWMA is a no-op EWMA.
 type NilEWMA struct{}
 
-// Rate is a no-op.
-func (NilEWMA) Rate() float64 { return 0.0 }
-
-// Snapshot is a no-op.
-func (NilEWMA) Snapshot() EWMA { return NilEWMA{} }
-
-// Tick is a no-op.
-func (NilEWMA) Tick() {}
-
-// Update is a no-op.
-func (NilEWMA) Update(n int64) {}
+func (NilEWMA) Snapshot() EWMASnapshot { return (*emptySnapshot)(nil) }
+func (NilEWMA) Tick()                  {}
+func (NilEWMA) Update(n int64)         {}
 
 // StandardEWMA is the standard implementation of an EWMA and tracks the number
 // of uncounted events and processes them on each tick.  It uses the
@@ -77,37 +59,50 @@ func (NilEWMA) Update(n int64) {}
 type StandardEWMA struct {
 	uncounted atomic.Int64
 	alpha     float64
-	rate      float64
-	init      bool
+	rate      atomic.Uint64
+	init      atomic.Bool
 	mutex     sync.Mutex
 }
 
-// Rate returns the moving average rate of events per second.
-func (a *StandardEWMA) Rate() float64 {
-	a.mutex.Lock()
-	defer a.mutex.Unlock()
-	return a.rate * float64(time.Second)
-}
-
 // Snapshot returns a read-only copy of the EWMA.
-func (a *StandardEWMA) Snapshot() EWMA {
-	return EWMASnapshot(a.Rate())
+func (a *StandardEWMA) Snapshot() EWMASnapshot {
+	r := math.Float64frombits(a.rate.Load()) * float64(time.Second)
+	return ewmaSnapshot(r)
 }
 
 // Tick ticks the clock to update the moving average.  It assumes it is called
 // every five seconds.
 func (a *StandardEWMA) Tick() {
-	count := a.uncounted.Load()
-	a.uncounted.Add(-count)
-	instantRate := float64(count) / float64(5*time.Second)
+	// Optimization to avoid mutex locking in the hot-path.
+	if a.init.Load() {
+		a.updateRate(a.fetchInstantRate())
+		return
+	}
+	// Slow-path: this is only needed on the first Tick() and preserves transactional updating
+	// of init and rate in the else block. The first conditional is needed below because
+	// a different thread could have set a.init = 1 between the time of the first atomic load and when
+	// the lock was acquired.
 	a.mutex.Lock()
-	defer a.mutex.Unlock()
-	if a.init {
-		a.rate += a.alpha * (instantRate - a.rate)
+	if a.init.Load() {
+		// The fetchInstantRate() uses atomic loading, which is unnecessary in this critical section
+		// but again, this section is only invoked on the first successful Tick() operation.
+		a.updateRate(a.fetchInstantRate())
 	} else {
-		a.init = true
-		a.rate = instantRate
+		a.init.Store(true)
+		a.rate.Store(math.Float64bits(a.fetchInstantRate()))
 	}
+	a.mutex.Unlock()
+}
+
+func (a *StandardEWMA) fetchInstantRate() float64 {
+	count := a.uncounted.Swap(0)
+	return float64(count) / float64(5*time.Second)
+}
+
+func (a *StandardEWMA) updateRate(instantRate float64) {
+	currentRate := math.Float64frombits(a.rate.Load())
+	currentRate += a.alpha * (instantRate - currentRate)
+	a.rate.Store(math.Float64bits(currentRate))
 }
 
 // Update adds n uncounted events.
diff --git a/metrics/ewma_test.go b/metrics/ewma_test.go
index 5b24419161..9a91b43db8 100644
--- a/metrics/ewma_test.go
+++ b/metrics/ewma_test.go
@@ -5,6 +5,8 @@ import (
 	"testing"
 )
 
+const epsilon = 0.0000000000000001
+
 func BenchmarkEWMA(b *testing.B) {
 	a := NewEWMA1()
 	b.ResetTimer()
@@ -14,72 +16,33 @@ func BenchmarkEWMA(b *testing.B) {
 	}
 }
 
+func BenchmarkEWMAParallel(b *testing.B) {
+	a := NewEWMA1()
+	b.ResetTimer()
+
+	b.RunParallel(func(pb *testing.PB) {
+		for pb.Next() {
+			a.Update(1)
+			a.Tick()
+		}
+	})
+}
+
 func TestEWMA1(t *testing.T) {
 	a := NewEWMA1()
 	a.Update(3)
 	a.Tick()
-	if rate := a.Rate(); math.Abs(0.6-rate) > epsilon {
-		t.Errorf("initial a.Rate(): 0.6 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.22072766470286553-rate) > epsilon {
-		t.Errorf("1 minute a.Rate(): 0.22072766470286553 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.08120116994196772-rate) > epsilon {
-		t.Errorf("2 minute a.Rate(): 0.08120116994196772 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.029872241020718428-rate) > epsilon {
-		t.Errorf("3 minute a.Rate(): 0.029872241020718428 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.01098938333324054-rate) > epsilon {
-		t.Errorf("4 minute a.Rate(): 0.01098938333324054 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.004042768199451294-rate) > epsilon {
-		t.Errorf("5 minute a.Rate(): 0.004042768199451294 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.0014872513059998212-rate) > epsilon {
-		t.Errorf("6 minute a.Rate(): 0.0014872513059998212 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.0005471291793327122-rate) > epsilon {
-		t.Errorf("7 minute a.Rate(): 0.0005471291793327122 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.00020127757674150815-rate) > epsilon {
-		t.Errorf("8 minute a.Rate(): 0.00020127757674150815 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(7.404588245200814e-05-rate) > epsilon {
-		t.Errorf("9 minute a.Rate(): 7.404588245200814e-05 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(2.7239957857491083e-05-rate) > epsilon {
-		t.Errorf("10 minute a.Rate(): 2.7239957857491083e-05 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(1.0021020474147462e-05-rate) > epsilon {
-		t.Errorf("11 minute a.Rate(): 1.0021020474147462e-05 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(3.6865274119969525e-06-rate) > epsilon {
-		t.Errorf("12 minute a.Rate(): 3.6865274119969525e-06 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(1.3561976441886433e-06-rate) > epsilon {
-		t.Errorf("13 minute a.Rate(): 1.3561976441886433e-06 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(4.989172314621449e-07-rate) > epsilon {
-		t.Errorf("14 minute a.Rate(): 4.989172314621449e-07 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(1.8354139230109722e-07-rate) > epsilon {
-		t.Errorf("15 minute a.Rate(): 1.8354139230109722e-07 != %v\n", rate)
+	for i, want := range []float64{0.6,
+		0.22072766470286553, 0.08120116994196772, 0.029872241020718428,
+		0.01098938333324054, 0.004042768199451294, 0.0014872513059998212,
+		0.0005471291793327122, 0.00020127757674150815, 7.404588245200814e-05,
+		2.7239957857491083e-05, 1.0021020474147462e-05, 3.6865274119969525e-06,
+		1.3561976441886433e-06, 4.989172314621449e-07, 1.8354139230109722e-07,
+	} {
+		if rate := a.Snapshot().Rate(); math.Abs(want-rate) > epsilon {
+			t.Errorf("%d minute a.Snapshot().Rate(): %f != %v\n", i, want, rate)
+		}
+		elapseMinute(a)
 	}
 }
 
@@ -87,68 +50,17 @@ func TestEWMA5(t *testing.T) {
 	a := NewEWMA5()
 	a.Update(3)
 	a.Tick()
-	if rate := a.Rate(); math.Abs(0.6-rate) > epsilon {
-		t.Errorf("initial a.Rate(): 0.6 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.49123845184678905-rate) > epsilon {
-		t.Errorf("1 minute a.Rate(): 0.49123845184678905 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.4021920276213837-rate) > epsilon {
-		t.Errorf("2 minute a.Rate(): 0.4021920276213837 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.32928698165641596-rate) > epsilon {
-		t.Errorf("3 minute a.Rate(): 0.32928698165641596 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.269597378470333-rate) > epsilon {
-		t.Errorf("4 minute a.Rate(): 0.269597378470333 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.2207276647028654-rate) > epsilon {
-		t.Errorf("5 minute a.Rate(): 0.2207276647028654 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.18071652714732128-rate) > epsilon {
-		t.Errorf("6 minute a.Rate(): 0.18071652714732128 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.14795817836496392-rate) > epsilon {
-		t.Errorf("7 minute a.Rate(): 0.14795817836496392 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.12113791079679326-rate) > epsilon {
-		t.Errorf("8 minute a.Rate(): 0.12113791079679326 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.09917933293295193-rate) > epsilon {
-		t.Errorf("9 minute a.Rate(): 0.09917933293295193 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.08120116994196763-rate) > epsilon {
-		t.Errorf("10 minute a.Rate(): 0.08120116994196763 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.06648189501740036-rate) > epsilon {
-		t.Errorf("11 minute a.Rate(): 0.06648189501740036 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.05443077197364752-rate) > epsilon {
-		t.Errorf("12 minute a.Rate(): 0.05443077197364752 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.04456414692860035-rate) > epsilon {
-		t.Errorf("13 minute a.Rate(): 0.04456414692860035 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.03648603757513079-rate) > epsilon {
-		t.Errorf("14 minute a.Rate(): 0.03648603757513079 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.0298722410207183831020718428-rate) > epsilon {
-		t.Errorf("15 minute a.Rate(): 0.0298722410207183831020718428 != %v\n", rate)
+	for i, want := range []float64{
+		0.6, 0.49123845184678905, 0.4021920276213837, 0.32928698165641596,
+		0.269597378470333, 0.2207276647028654, 0.18071652714732128,
+		0.14795817836496392, 0.12113791079679326, 0.09917933293295193,
+		0.08120116994196763, 0.06648189501740036, 0.05443077197364752,
+		0.04456414692860035, 0.03648603757513079, 0.0298722410207183831020718428,
+	} {
+		if rate := a.Snapshot().Rate(); math.Abs(want-rate) > epsilon {
+			t.Errorf("%d minute a.Snapshot().Rate(): %f != %v\n", i, want, rate)
+		}
+		elapseMinute(a)
 	}
 }
 
@@ -156,68 +68,17 @@ func TestEWMA15(t *testing.T) {
 	a := NewEWMA15()
 	a.Update(3)
 	a.Tick()
-	if rate := a.Rate(); math.Abs(0.6-rate) > epsilon {
-		t.Errorf("initial a.Rate(): 0.6 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.5613041910189706-rate) > epsilon {
-		t.Errorf("1 minute a.Rate(): 0.5613041910189706 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.5251039914257684-rate) > epsilon {
-		t.Errorf("2 minute a.Rate(): 0.5251039914257684 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.4912384518467888184678905-rate) > epsilon {
-		t.Errorf("3 minute a.Rate(): 0.4912384518467888184678905 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.459557003018789-rate) > epsilon {
-		t.Errorf("4 minute a.Rate(): 0.459557003018789 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.4299187863442732-rate) > epsilon {
-		t.Errorf("5 minute a.Rate(): 0.4299187863442732 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.4021920276213831-rate) > epsilon {
-		t.Errorf("6 minute a.Rate(): 0.4021920276213831 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.37625345116383313-rate) > epsilon {
-		t.Errorf("7 minute a.Rate(): 0.37625345116383313 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.3519877317060185-rate) > epsilon {
-		t.Errorf("8 minute a.Rate(): 0.3519877317060185 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.3292869816564153165641596-rate) > epsilon {
-		t.Errorf("9 minute a.Rate(): 0.3292869816564153165641596 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.3080502714195546-rate) > epsilon {
-		t.Errorf("10 minute a.Rate(): 0.3080502714195546 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.2881831806538789-rate) > epsilon {
-		t.Errorf("11 minute a.Rate(): 0.2881831806538789 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.26959737847033216-rate) > epsilon {
-		t.Errorf("12 minute a.Rate(): 0.26959737847033216 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.2522102307052083-rate) > epsilon {
-		t.Errorf("13 minute a.Rate(): 0.2522102307052083 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.23594443252115815-rate) > epsilon {
-		t.Errorf("14 minute a.Rate(): 0.23594443252115815 != %v\n", rate)
-	}
-	elapseMinute(a)
-	if rate := a.Rate(); math.Abs(0.2207276647028646247028654470286553-rate) > epsilon {
-		t.Errorf("15 minute a.Rate(): 0.2207276647028646247028654470286553 != %v\n", rate)
+	for i, want := range []float64{
+		0.6, 0.5613041910189706, 0.5251039914257684, 0.4912384518467888184678905,
+		0.459557003018789, 0.4299187863442732, 0.4021920276213831,
+		0.37625345116383313, 0.3519877317060185, 0.3292869816564153165641596,
+		0.3080502714195546, 0.2881831806538789, 0.26959737847033216,
+		0.2522102307052083, 0.23594443252115815, 0.2207276647028646247028654470286553,
+	} {
+		if rate := a.Snapshot().Rate(); math.Abs(want-rate) > epsilon {
+			t.Errorf("%d minute a.Snapshot().Rate(): %f != %v\n", i, want, rate)
+		}
+		elapseMinute(a)
 	}
 }
 
diff --git a/metrics/exp/exp.go b/metrics/exp/exp.go
index 9e850f96b2..7e3f82a075 100644
--- a/metrics/exp/exp.go
+++ b/metrics/exp/exp.go
@@ -109,25 +109,25 @@ func (exp *exp) getInfo(name string) *expvar.String {
 	return v
 }
 
-func (exp *exp) publishCounter(name string, metric metrics.Counter) {
+func (exp *exp) publishCounter(name string, metric metrics.CounterSnapshot) {
 	v := exp.getInt(name)
 	v.Set(metric.Count())
 }
 
-func (exp *exp) publishCounterFloat64(name string, metric metrics.CounterFloat64) {
+func (exp *exp) publishCounterFloat64(name string, metric metrics.CounterFloat64Snapshot) {
 	v := exp.getFloat(name)
 	v.Set(metric.Count())
 }
 
-func (exp *exp) publishGauge(name string, metric metrics.Gauge) {
+func (exp *exp) publishGauge(name string, metric metrics.GaugeSnapshot) {
 	v := exp.getInt(name)
 	v.Set(metric.Value())
 }
-func (exp *exp) publishGaugeFloat64(name string, metric metrics.GaugeFloat64) {
+func (exp *exp) publishGaugeFloat64(name string, metric metrics.GaugeFloat64Snapshot) {
 	exp.getFloat(name).Set(metric.Value())
 }
 
-func (exp *exp) publishGaugeInfo(name string, metric metrics.GaugeInfo) {
+func (exp *exp) publishGaugeInfo(name string, metric metrics.GaugeInfoSnapshot) {
 	exp.getInfo(name).Set(metric.Value().String())
 }
 
@@ -176,28 +176,28 @@ func (exp *exp) publishTimer(name string, metric metrics.Timer) {
 
 func (exp *exp) publishResettingTimer(name string, metric metrics.ResettingTimer) {
 	t := metric.Snapshot()
-	ps := t.Percentiles([]float64{50, 75, 95, 99})
-	exp.getInt(name + ".count").Set(int64(len(t.Values())))
+	ps := t.Percentiles([]float64{0.50, 0.75, 0.95, 0.99})
+	exp.getInt(name + ".count").Set(int64(t.Count()))
 	exp.getFloat(name + ".mean").Set(t.Mean())
-	exp.getInt(name + ".50-percentile").Set(ps[0])
-	exp.getInt(name + ".75-percentile").Set(ps[1])
-	exp.getInt(name + ".95-percentile").Set(ps[2])
-	exp.getInt(name + ".99-percentile").Set(ps[3])
+	exp.getFloat(name + ".50-percentile").Set(ps[0])
+	exp.getFloat(name + ".75-percentile").Set(ps[1])
+	exp.getFloat(name + ".95-percentile").Set(ps[2])
+	exp.getFloat(name + ".99-percentile").Set(ps[3])
 }
 
 func (exp *exp) syncToExpvar() {
 	exp.registry.Each(func(name string, i interface{}) {
 		switch i := i.(type) {
 		case metrics.Counter:
-			exp.publishCounter(name, i)
+			exp.publishCounter(name, i.Snapshot())
 		case metrics.CounterFloat64:
-			exp.publishCounterFloat64(name, i)
+			exp.publishCounterFloat64(name, i.Snapshot())
 		case metrics.Gauge:
-			exp.publishGauge(name, i)
+			exp.publishGauge(name, i.Snapshot())
 		case metrics.GaugeFloat64:
-			exp.publishGaugeFloat64(name, i)
+			exp.publishGaugeFloat64(name, i.Snapshot())
 		case metrics.GaugeInfo:
-			exp.publishGaugeInfo(name, i)
+			exp.publishGaugeInfo(name, i.Snapshot())
 		case metrics.Histogram:
 			exp.publishHistogram(name, i)
 		case metrics.Meter:
diff --git a/metrics/gauge.go b/metrics/gauge.go
index 81137d7f7c..68f8f11abc 100644
--- a/metrics/gauge.go
+++ b/metrics/gauge.go
@@ -2,13 +2,18 @@ package metrics
 
 import "sync/atomic"
 
+// gaugeSnapshot contains a readonly int64.
+type GaugeSnapshot interface {
+	Value() int64
+}
+
 // Gauges hold an int64 value that can be set arbitrarily.
 type Gauge interface {
-	Snapshot() Gauge
+	Snapshot() GaugeSnapshot
 	Update(int64)
+	UpdateIfGt(int64)
 	Dec(int64)
 	Inc(int64)
-	Value() int64
 }
 
 // GetOrRegisterGauge returns an existing Gauge or constructs and registers a
@@ -38,65 +43,20 @@ func NewRegisteredGauge(name string, r Registry) Gauge {
 	return c
 }
 
-// NewFunctionalGauge constructs a new FunctionalGauge.
-func NewFunctionalGauge(f func() int64) Gauge {
-	if !Enabled {
-		return NilGauge{}
-	}
-	return &FunctionalGauge{value: f}
-}
-
-// NewRegisteredFunctionalGauge constructs and registers a new StandardGauge.
-func NewRegisteredFunctionalGauge(name string, r Registry, f func() int64) Gauge {
-	c := NewFunctionalGauge(f)
-	if nil == r {
-		r = DefaultRegistry
-	}
-	r.Register(name, c)
-	return c
-}
-
-// GaugeSnapshot is a read-only copy of another Gauge.
-type GaugeSnapshot int64
-
-// Snapshot returns the snapshot.
-func (g GaugeSnapshot) Snapshot() Gauge { return g }
-
-// Update panics.
-func (GaugeSnapshot) Update(int64) {
-	panic("Update called on a GaugeSnapshot")
-}
-
-// Dec panics.
-func (GaugeSnapshot) Dec(int64) {
-	panic("Dec called on a GaugeSnapshot")
-}
-
-// Inc panics.
-func (GaugeSnapshot) Inc(int64) {
-	panic("Inc called on a GaugeSnapshot")
-}
+// gaugeSnapshot is a read-only copy of another Gauge.
+type gaugeSnapshot int64
 
 // Value returns the value at the time the snapshot was taken.
-func (g GaugeSnapshot) Value() int64 { return int64(g) }
+func (g gaugeSnapshot) Value() int64 { return int64(g) }
 
 // NilGauge is a no-op Gauge.
 type NilGauge struct{}
 
-// Snapshot is a no-op.
-func (NilGauge) Snapshot() Gauge { return NilGauge{} }
-
-// Update is a no-op.
-func (NilGauge) Update(v int64) {}
-
-// Dec is a no-op.
-func (NilGauge) Dec(i int64) {}
-
-// Inc is a no-op.
-func (NilGauge) Inc(i int64) {}
-
-// Value is a no-op.
-func (NilGauge) Value() int64 { return 0 }
+func (NilGauge) Snapshot() GaugeSnapshot { return (*emptySnapshot)(nil) }
+func (NilGauge) Update(v int64)          {}
+func (NilGauge) UpdateIfGt(v int64)      {}
+func (NilGauge) Dec(i int64)             {}
+func (NilGauge) Inc(i int64)             {}
 
 // StandardGauge is the standard implementation of a Gauge and uses the
 // sync/atomic package to manage a single int64 value.
@@ -105,8 +65,8 @@ type StandardGauge struct {
 }
 
 // Snapshot returns a read-only copy of the gauge.
-func (g *StandardGauge) Snapshot() Gauge {
-	return GaugeSnapshot(g.Value())
+func (g *StandardGauge) Snapshot() GaugeSnapshot {
+	return gaugeSnapshot(g.value.Load())
 }
 
 // Update updates the gauge's value.
@@ -114,9 +74,17 @@ func (g *StandardGauge) Update(v int64) {
 	g.value.Store(v)
 }
 
-// Value returns the gauge's current value.
-func (g *StandardGauge) Value() int64 {
-	return g.value.Load()
+// Update updates the gauge's value if v is larger then the current valie.
+func (g *StandardGauge) UpdateIfGt(v int64) {
+	for {
+		exist := g.value.Load()
+		if exist >= v {
+			break
+		}
+		if g.value.CompareAndSwap(exist, v) {
+			break
+		}
+	}
 }
 
 // Dec decrements the gauge's current value by the given amount.
@@ -128,31 +96,3 @@ func (g *StandardGauge) Dec(i int64) {
 func (g *StandardGauge) Inc(i int64) {
 	g.value.Add(i)
 }
-
-// FunctionalGauge returns value from given function
-type FunctionalGauge struct {
-	value func() int64
-}
-
-// Value returns the gauge's current value.
-func (g FunctionalGauge) Value() int64 {
-	return g.value()
-}
-
-// Snapshot returns the snapshot.
-func (g FunctionalGauge) Snapshot() Gauge { return GaugeSnapshot(g.Value()) }
-
-// Update panics.
-func (FunctionalGauge) Update(int64) {
-	panic("Update called on a FunctionalGauge")
-}
-
-// Dec panics.
-func (FunctionalGauge) Dec(int64) {
-	panic("Dec called on a FunctionalGauge")
-}
-
-// Inc panics.
-func (FunctionalGauge) Inc(int64) {
-	panic("Inc called on a FunctionalGauge")
-}
diff --git a/metrics/gauge_float64.go b/metrics/gauge_float64.go
index 237ff8036e..967f2bc60e 100644
--- a/metrics/gauge_float64.go
+++ b/metrics/gauge_float64.go
@@ -5,11 +5,14 @@ import (
 	"sync/atomic"
 )
 
-// GaugeFloat64s hold a float64 value that can be set arbitrarily.
+type GaugeFloat64Snapshot interface {
+	Value() float64
+}
+
+// GaugeFloat64 hold a float64 value that can be set arbitrarily.
 type GaugeFloat64 interface {
-	Snapshot() GaugeFloat64
+	Snapshot() GaugeFloat64Snapshot
 	Update(float64)
-	Value() float64
 }
 
 // GetOrRegisterGaugeFloat64 returns an existing GaugeFloat64 or constructs and registers a
@@ -39,49 +42,18 @@ func NewRegisteredGaugeFloat64(name string, r Registry) GaugeFloat64 {
 	return c
 }
 
-// NewFunctionalGauge constructs a new FunctionalGauge.
-func NewFunctionalGaugeFloat64(f func() float64) GaugeFloat64 {
-	if !Enabled {
-		return NilGaugeFloat64{}
-	}
-	return &FunctionalGaugeFloat64{value: f}
-}
-
-// NewRegisteredFunctionalGauge constructs and registers a new StandardGauge.
-func NewRegisteredFunctionalGaugeFloat64(name string, r Registry, f func() float64) GaugeFloat64 {
-	c := NewFunctionalGaugeFloat64(f)
-	if nil == r {
-		r = DefaultRegistry
-	}
-	r.Register(name, c)
-	return c
-}
-
-// GaugeFloat64Snapshot is a read-only copy of another GaugeFloat64.
-type GaugeFloat64Snapshot float64
-
-// Snapshot returns the snapshot.
-func (g GaugeFloat64Snapshot) Snapshot() GaugeFloat64 { return g }
-
-// Update panics.
-func (GaugeFloat64Snapshot) Update(float64) {
-	panic("Update called on a GaugeFloat64Snapshot")
-}
+// gaugeFloat64Snapshot is a read-only copy of another GaugeFloat64.
+type gaugeFloat64Snapshot float64
 
 // Value returns the value at the time the snapshot was taken.
-func (g GaugeFloat64Snapshot) Value() float64 { return float64(g) }
+func (g gaugeFloat64Snapshot) Value() float64 { return float64(g) }
 
 // NilGauge is a no-op Gauge.
 type NilGaugeFloat64 struct{}
 
-// Snapshot is a no-op.
-func (NilGaugeFloat64) Snapshot() GaugeFloat64 { return NilGaugeFloat64{} }
-
-// Update is a no-op.
-func (NilGaugeFloat64) Update(v float64) {}
-
-// Value is a no-op.
-func (NilGaugeFloat64) Value() float64 { return 0.0 }
+func (NilGaugeFloat64) Snapshot() GaugeFloat64Snapshot { return NilGaugeFloat64{} }
+func (NilGaugeFloat64) Update(v float64)               {}
+func (NilGaugeFloat64) Value() float64                 { return 0.0 }
 
 // StandardGaugeFloat64 is the standard implementation of a GaugeFloat64 and uses
 // atomic to manage a single float64 value.
@@ -90,34 +62,12 @@ type StandardGaugeFloat64 struct {
 }
 
 // Snapshot returns a read-only copy of the gauge.
-func (g *StandardGaugeFloat64) Snapshot() GaugeFloat64 {
-	return GaugeFloat64Snapshot(g.Value())
+func (g *StandardGaugeFloat64) Snapshot() GaugeFloat64Snapshot {
+	v := math.Float64frombits(g.floatBits.Load())
+	return gaugeFloat64Snapshot(v)
 }
 
 // Update updates the gauge's value.
 func (g *StandardGaugeFloat64) Update(v float64) {
 	g.floatBits.Store(math.Float64bits(v))
 }
-
-// Value returns the gauge's current value.
-func (g *StandardGaugeFloat64) Value() float64 {
-	return math.Float64frombits(g.floatBits.Load())
-}
-
-// FunctionalGaugeFloat64 returns value from given function
-type FunctionalGaugeFloat64 struct {
-	value func() float64
-}
-
-// Value returns the gauge's current value.
-func (g FunctionalGaugeFloat64) Value() float64 {
-	return g.value()
-}
-
-// Snapshot returns the snapshot.
-func (g FunctionalGaugeFloat64) Snapshot() GaugeFloat64 { return GaugeFloat64Snapshot(g.Value()) }
-
-// Update panics.
-func (FunctionalGaugeFloat64) Update(float64) {
-	panic("Update called on a FunctionalGaugeFloat64")
-}
diff --git a/metrics/gauge_float64_test.go b/metrics/gauge_float64_test.go
index 647d090009..f0ac7ea5e7 100644
--- a/metrics/gauge_float64_test.go
+++ b/metrics/gauge_float64_test.go
@@ -26,19 +26,11 @@ func BenchmarkGaugeFloat64Parallel(b *testing.B) {
 		}()
 	}
 	wg.Wait()
-	if have, want := c.Value(), float64(b.N-1); have != want {
+	if have, want := c.Snapshot().Value(), float64(b.N-1); have != want {
 		b.Fatalf("have %f want %f", have, want)
 	}
 }
 
-func TestGaugeFloat64(t *testing.T) {
-	g := NewGaugeFloat64()
-	g.Update(47.0)
-	if v := g.Value(); 47.0 != v {
-		t.Errorf("g.Value(): 47.0 != %v\n", v)
-	}
-}
-
 func TestGaugeFloat64Snapshot(t *testing.T) {
 	g := NewGaugeFloat64()
 	g.Update(47.0)
@@ -53,28 +45,7 @@ func TestGetOrRegisterGaugeFloat64(t *testing.T) {
 	r := NewRegistry()
 	NewRegisteredGaugeFloat64("foo", r).Update(47.0)
 	t.Logf("registry: %v", r)
-	if g := GetOrRegisterGaugeFloat64("foo", r); 47.0 != g.Value() {
-		t.Fatal(g)
-	}
-}
-
-func TestFunctionalGaugeFloat64(t *testing.T) {
-	var counter float64
-	fg := NewFunctionalGaugeFloat64(func() float64 {
-		counter++
-		return counter
-	})
-	fg.Value()
-	fg.Value()
-	if counter != 2 {
-		t.Error("counter != 2")
-	}
-}
-
-func TestGetOrRegisterFunctionalGaugeFloat64(t *testing.T) {
-	r := NewRegistry()
-	NewRegisteredFunctionalGaugeFloat64("foo", r, func() float64 { return 47 })
-	if g := GetOrRegisterGaugeFloat64("foo", r); g.Value() != 47 {
+	if g := GetOrRegisterGaugeFloat64("foo", r).Snapshot(); 47.0 != g.Value() {
 		t.Fatal(g)
 	}
 }
diff --git a/metrics/gauge_info.go b/metrics/gauge_info.go
index f1b2075939..c44b2d85f3 100644
--- a/metrics/gauge_info.go
+++ b/metrics/gauge_info.go
@@ -5,14 +5,17 @@ import (
 	"sync"
 )
 
+type GaugeInfoSnapshot interface {
+	Value() GaugeInfoValue
+}
+
 // GaugeInfos hold a GaugeInfoValue value that can be set arbitrarily.
 type GaugeInfo interface {
-	Snapshot() GaugeInfo
 	Update(GaugeInfoValue)
-	Value() GaugeInfoValue
+	Snapshot() GaugeInfoSnapshot
 }
 
-// GaugeInfoValue is a mappng of (string) keys to (string) values
+// GaugeInfoValue is a mapping of keys to values
 type GaugeInfoValue map[string]string
 
 func (val GaugeInfoValue) String() string {
@@ -49,49 +52,17 @@ func NewRegisteredGaugeInfo(name string, r Registry) GaugeInfo {
 	return c
 }
 
-// NewFunctionalGauge constructs a new FunctionalGauge.
-func NewFunctionalGaugeInfo(f func() GaugeInfoValue) GaugeInfo {
-	if !Enabled {
-		return NilGaugeInfo{}
-	}
-	return &FunctionalGaugeInfo{value: f}
-}
-
-// NewRegisteredFunctionalGauge constructs and registers a new StandardGauge.
-func NewRegisteredFunctionalGaugeInfo(name string, r Registry, f func() GaugeInfoValue) GaugeInfo {
-	c := NewFunctionalGaugeInfo(f)
-	if nil == r {
-		r = DefaultRegistry
-	}
-	r.Register(name, c)
-	return c
-}
-
-// GaugeInfoSnapshot is a read-only copy of another GaugeInfo.
-type GaugeInfoSnapshot GaugeInfoValue
-
-// Snapshot returns the snapshot.
-func (g GaugeInfoSnapshot) Snapshot() GaugeInfo { return g }
-
-// Update panics.
-func (GaugeInfoSnapshot) Update(GaugeInfoValue) {
-	panic("Update called on a GaugeInfoSnapshot")
-}
+// gaugeInfoSnapshot is a read-only copy of another GaugeInfo.
+type gaugeInfoSnapshot GaugeInfoValue
 
 // Value returns the value at the time the snapshot was taken.
-func (g GaugeInfoSnapshot) Value() GaugeInfoValue { return GaugeInfoValue(g) }
+func (g gaugeInfoSnapshot) Value() GaugeInfoValue { return GaugeInfoValue(g) }
 
-// NilGauge is a no-op Gauge.
 type NilGaugeInfo struct{}
 
-// Snapshot is a no-op.
-func (NilGaugeInfo) Snapshot() GaugeInfo { return NilGaugeInfo{} }
-
-// Update is a no-op.
-func (NilGaugeInfo) Update(v GaugeInfoValue) {}
-
-// Value is a no-op.
-func (NilGaugeInfo) Value() GaugeInfoValue { return GaugeInfoValue{} }
+func (NilGaugeInfo) Snapshot() GaugeInfoSnapshot { return NilGaugeInfo{} }
+func (NilGaugeInfo) Update(v GaugeInfoValue)     {}
+func (NilGaugeInfo) Value() GaugeInfoValue       { return GaugeInfoValue{} }
 
 // StandardGaugeInfo is the standard implementation of a GaugeInfo and uses
 // sync.Mutex to manage a single string value.
@@ -101,8 +72,8 @@ type StandardGaugeInfo struct {
 }
 
 // Snapshot returns a read-only copy of the gauge.
-func (g *StandardGaugeInfo) Snapshot() GaugeInfo {
-	return GaugeInfoSnapshot(g.Value())
+func (g *StandardGaugeInfo) Snapshot() GaugeInfoSnapshot {
+	return gaugeInfoSnapshot(g.value)
 }
 
 // Update updates the gauge's value.
@@ -111,34 +82,3 @@ func (g *StandardGaugeInfo) Update(v GaugeInfoValue) {
 	defer g.mutex.Unlock()
 	g.value = v
 }
-
-// Value returns the gauge's current value.
-func (g *StandardGaugeInfo) Value() GaugeInfoValue {
-	g.mutex.Lock()
-	defer g.mutex.Unlock()
-	return g.value
-}
-
-// FunctionalGaugeInfo returns value from given function
-type FunctionalGaugeInfo struct {
-	value func() GaugeInfoValue
-}
-
-// Value returns the gauge's current value.
-func (g FunctionalGaugeInfo) Value() GaugeInfoValue {
-	return g.value()
-}
-
-// Value returns the gauge's current value in JSON string format
-func (g FunctionalGaugeInfo) ValueJsonString() string {
-	data, _ := json.Marshal(g.value())
-	return string(data)
-}
-
-// Snapshot returns the snapshot.
-func (g FunctionalGaugeInfo) Snapshot() GaugeInfo { return GaugeInfoSnapshot(g.Value()) }
-
-// Update panics.
-func (FunctionalGaugeInfo) Update(GaugeInfoValue) {
-	panic("Update called on a FunctionalGaugeInfo")
-}
diff --git a/metrics/gauge_info_test.go b/metrics/gauge_info_test.go
index 4227a6a85f..319afbf92e 100644
--- a/metrics/gauge_info_test.go
+++ b/metrics/gauge_info_test.go
@@ -1,7 +1,6 @@
 package metrics
 
 import (
-	"strconv"
 	"testing"
 )
 
@@ -14,22 +13,14 @@ func TestGaugeInfoJsonString(t *testing.T) {
 	},
 	)
 	want := `{"anotherKey":"any_string_value","chain_id":"5","third_key":"anything"}`
-	if have := g.Value().String(); have != want {
-		t.Errorf("\nhave: %v\nwant: %v\n", have, want)
-	}
-}
 
-func TestGaugeInfoSnapshot(t *testing.T) {
-	g := NewGaugeInfo()
-	g.Update(GaugeInfoValue{"value": "original"})
-	snapshot := g.Snapshot() // Snapshot @chainid 5
+	original := g.Snapshot()
 	g.Update(GaugeInfoValue{"value": "updated"})
-	// The 'g' should be updated
-	if have, want := g.Value().String(), `{"value":"updated"}`; have != want {
+
+	if have := original.Value().String(); have != want {
 		t.Errorf("\nhave: %v\nwant: %v\n", have, want)
 	}
-	// Snapshot should be unupdated
-	if have, want := snapshot.Value().String(), `{"value":"original"}`; have != want {
+	if have, want := g.Snapshot().Value().String(), `{"value":"updated"}`; have != want {
 		t.Errorf("\nhave: %v\nwant: %v\n", have, want)
 	}
 }
@@ -38,38 +29,8 @@ func TestGetOrRegisterGaugeInfo(t *testing.T) {
 	r := NewRegistry()
 	NewRegisteredGaugeInfo("foo", r).Update(
 		GaugeInfoValue{"chain_id": "5"})
-	g := GetOrRegisterGaugeInfo("foo", r)
+	g := GetOrRegisterGaugeInfo("foo", r).Snapshot()
 	if have, want := g.Value().String(), `{"chain_id":"5"}`; have != want {
 		t.Errorf("have\n%v\nwant\n%v\n", have, want)
 	}
 }
-
-func TestFunctionalGaugeInfo(t *testing.T) {
-	info := GaugeInfoValue{"chain_id": "0"}
-	counter := 1
-	// A "functional" gauge invokes the method to obtain the value
-	fg := NewFunctionalGaugeInfo(func() GaugeInfoValue {
-		info["chain_id"] = strconv.Itoa(counter)
-		counter++
-		return info
-	})
-	fg.Value()
-	fg.Value()
-	if have, want := info["chain_id"], "2"; have != want {
-		t.Errorf("have %v want %v", have, want)
-	}
-}
-
-func TestGetOrRegisterFunctionalGaugeInfo(t *testing.T) {
-	r := NewRegistry()
-	NewRegisteredFunctionalGaugeInfo("foo", r, func() GaugeInfoValue {
-		return GaugeInfoValue{
-			"chain_id": "5",
-		}
-	})
-	want := `{"chain_id":"5"}`
-	have := GetOrRegisterGaugeInfo("foo", r).Value().String()
-	if have != want {
-		t.Errorf("have\n%v\nwant\n%v\n", have, want)
-	}
-}
diff --git a/metrics/gauge_test.go b/metrics/gauge_test.go
index a98fe985d8..f2ba930bc4 100644
--- a/metrics/gauge_test.go
+++ b/metrics/gauge_test.go
@@ -1,7 +1,6 @@
 package metrics
 
 import (
-	"fmt"
 	"testing"
 )
 
@@ -13,14 +12,6 @@ func BenchmarkGauge(b *testing.B) {
 	}
 }
 
-func TestGauge(t *testing.T) {
-	g := NewGauge()
-	g.Update(int64(47))
-	if v := g.Value(); v != 47 {
-		t.Errorf("g.Value(): 47 != %v\n", v)
-	}
-}
-
 func TestGaugeSnapshot(t *testing.T) {
 	g := NewGauge()
 	g.Update(int64(47))
@@ -34,35 +25,7 @@ func TestGaugeSnapshot(t *testing.T) {
 func TestGetOrRegisterGauge(t *testing.T) {
 	r := NewRegistry()
 	NewRegisteredGauge("foo", r).Update(47)
-	if g := GetOrRegisterGauge("foo", r); g.Value() != 47 {
-		t.Fatal(g)
-	}
-}
-
-func TestFunctionalGauge(t *testing.T) {
-	var counter int64
-	fg := NewFunctionalGauge(func() int64 {
-		counter++
-		return counter
-	})
-	fg.Value()
-	fg.Value()
-	if counter != 2 {
-		t.Error("counter != 2")
-	}
-}
-
-func TestGetOrRegisterFunctionalGauge(t *testing.T) {
-	r := NewRegistry()
-	NewRegisteredFunctionalGauge("foo", r, func() int64 { return 47 })
-	if g := GetOrRegisterGauge("foo", r); g.Value() != 47 {
+	if g := GetOrRegisterGauge("foo", r); g.Snapshot().Value() != 47 {
 		t.Fatal(g)
 	}
 }
-
-func ExampleGetOrRegisterGauge() {
-	m := "server.bytes_sent"
-	g := GetOrRegisterGauge(m, nil)
-	g.Update(47)
-	fmt.Println(g.Value()) // Output: 47
-}
diff --git a/metrics/graphite.go b/metrics/graphite.go
index 4e3dd3b3b8..aba752e0ed 100644
--- a/metrics/graphite.go
+++ b/metrics/graphite.go
@@ -66,15 +66,15 @@ func graphite(c *GraphiteConfig) error {
 	c.Registry.Each(func(name string, i interface{}) {
 		switch metric := i.(type) {
 		case Counter:
-			fmt.Fprintf(w, "%s.%s.count %d %d\n", c.Prefix, name, metric.Count(), now)
+			fmt.Fprintf(w, "%s.%s.count %d %d\n", c.Prefix, name, metric.Snapshot().Count(), now)
 		case CounterFloat64:
-			fmt.Fprintf(w, "%s.%s.count %f %d\n", c.Prefix, name, metric.Count(), now)
+			fmt.Fprintf(w, "%s.%s.count %f %d\n", c.Prefix, name, metric.Snapshot().Count(), now)
 		case Gauge:
-			fmt.Fprintf(w, "%s.%s.value %d %d\n", c.Prefix, name, metric.Value(), now)
+			fmt.Fprintf(w, "%s.%s.value %d %d\n", c.Prefix, name, metric.Snapshot().Value(), now)
 		case GaugeFloat64:
-			fmt.Fprintf(w, "%s.%s.value %f %d\n", c.Prefix, name, metric.Value(), now)
+			fmt.Fprintf(w, "%s.%s.value %f %d\n", c.Prefix, name, metric.Snapshot().Value(), now)
 		case GaugeInfo:
-			fmt.Fprintf(w, "%s.%s.value %s %d\n", c.Prefix, name, metric.Value().String(), now)
+			fmt.Fprintf(w, "%s.%s.value %s %d\n", c.Prefix, name, metric.Snapshot().Value().String(), now)
 		case Histogram:
 			h := metric.Snapshot()
 			ps := h.Percentiles(c.Percentiles)
diff --git a/metrics/histogram.go b/metrics/histogram.go
index 2c54ce8b40..44de588bc1 100644
--- a/metrics/histogram.go
+++ b/metrics/histogram.go
@@ -1,20 +1,14 @@
 package metrics
 
+type HistogramSnapshot interface {
+	SampleSnapshot
+}
+
 // Histograms calculate distribution statistics from a series of int64 values.
 type Histogram interface {
 	Clear()
-	Count() int64
-	Max() int64
-	Mean() float64
-	Min() int64
-	Percentile(float64) float64
-	Percentiles([]float64) []float64
-	Sample() Sample
-	Snapshot() Histogram
-	StdDev() float64
-	Sum() int64
 	Update(int64)
-	Variance() float64
+	Snapshot() HistogramSnapshot
 }
 
 // GetOrRegisterHistogram returns an existing Histogram or constructs and
@@ -54,108 +48,12 @@ func NewRegisteredHistogram(name string, r Registry, s Sample) Histogram {
 	return c
 }
 
-// HistogramSnapshot is a read-only copy of another Histogram.
-type HistogramSnapshot struct {
-	sample *SampleSnapshot
-}
-
-// Clear panics.
-func (*HistogramSnapshot) Clear() {
-	panic("Clear called on a HistogramSnapshot")
-}
-
-// Count returns the number of samples recorded at the time the snapshot was
-// taken.
-func (h *HistogramSnapshot) Count() int64 { return h.sample.Count() }
-
-// Max returns the maximum value in the sample at the time the snapshot was
-// taken.
-func (h *HistogramSnapshot) Max() int64 { return h.sample.Max() }
-
-// Mean returns the mean of the values in the sample at the time the snapshot
-// was taken.
-func (h *HistogramSnapshot) Mean() float64 { return h.sample.Mean() }
-
-// Min returns the minimum value in the sample at the time the snapshot was
-// taken.
-func (h *HistogramSnapshot) Min() int64 { return h.sample.Min() }
-
-// Percentile returns an arbitrary percentile of values in the sample at the
-// time the snapshot was taken.
-func (h *HistogramSnapshot) Percentile(p float64) float64 {
-	return h.sample.Percentile(p)
-}
-
-// Percentiles returns a slice of arbitrary percentiles of values in the sample
-// at the time the snapshot was taken.
-func (h *HistogramSnapshot) Percentiles(ps []float64) []float64 {
-	return h.sample.Percentiles(ps)
-}
-
-// Sample returns the Sample underlying the histogram.
-func (h *HistogramSnapshot) Sample() Sample { return h.sample }
-
-// Snapshot returns the snapshot.
-func (h *HistogramSnapshot) Snapshot() Histogram { return h }
-
-// StdDev returns the standard deviation of the values in the sample at the
-// time the snapshot was taken.
-func (h *HistogramSnapshot) StdDev() float64 { return h.sample.StdDev() }
-
-// Sum returns the sum in the sample at the time the snapshot was taken.
-func (h *HistogramSnapshot) Sum() int64 { return h.sample.Sum() }
-
-// Update panics.
-func (*HistogramSnapshot) Update(int64) {
-	panic("Update called on a HistogramSnapshot")
-}
-
-// Variance returns the variance of inputs at the time the snapshot was taken.
-func (h *HistogramSnapshot) Variance() float64 { return h.sample.Variance() }
-
 // NilHistogram is a no-op Histogram.
 type NilHistogram struct{}
 
-// Clear is a no-op.
-func (NilHistogram) Clear() {}
-
-// Count is a no-op.
-func (NilHistogram) Count() int64 { return 0 }
-
-// Max is a no-op.
-func (NilHistogram) Max() int64 { return 0 }
-
-// Mean is a no-op.
-func (NilHistogram) Mean() float64 { return 0.0 }
-
-// Min is a no-op.
-func (NilHistogram) Min() int64 { return 0 }
-
-// Percentile is a no-op.
-func (NilHistogram) Percentile(p float64) float64 { return 0.0 }
-
-// Percentiles is a no-op.
-func (NilHistogram) Percentiles(ps []float64) []float64 {
-	return make([]float64, len(ps))
-}
-
-// Sample is a no-op.
-func (NilHistogram) Sample() Sample { return NilSample{} }
-
-// Snapshot is a no-op.
-func (NilHistogram) Snapshot() Histogram { return NilHistogram{} }
-
-// StdDev is a no-op.
-func (NilHistogram) StdDev() float64 { return 0.0 }
-
-// Sum is a no-op.
-func (NilHistogram) Sum() int64 { return 0 }
-
-// Update is a no-op.
-func (NilHistogram) Update(v int64) {}
-
-// Variance is a no-op.
-func (NilHistogram) Variance() float64 { return 0.0 }
+func (NilHistogram) Clear()                      {}
+func (NilHistogram) Snapshot() HistogramSnapshot { return (*emptySnapshot)(nil) }
+func (NilHistogram) Update(v int64)              {}
 
 // StandardHistogram is the standard implementation of a Histogram and uses a
 // Sample to bound its memory use.
@@ -166,46 +64,10 @@ type StandardHistogram struct {
 // Clear clears the histogram and its sample.
 func (h *StandardHistogram) Clear() { h.sample.Clear() }
 
-// Count returns the number of samples recorded since the histogram was last
-// cleared.
-func (h *StandardHistogram) Count() int64 { return h.sample.Count() }
-
-// Max returns the maximum value in the sample.
-func (h *StandardHistogram) Max() int64 { return h.sample.Max() }
-
-// Mean returns the mean of the values in the sample.
-func (h *StandardHistogram) Mean() float64 { return h.sample.Mean() }
-
-// Min returns the minimum value in the sample.
-func (h *StandardHistogram) Min() int64 { return h.sample.Min() }
-
-// Percentile returns an arbitrary percentile of the values in the sample.
-func (h *StandardHistogram) Percentile(p float64) float64 {
-	return h.sample.Percentile(p)
-}
-
-// Percentiles returns a slice of arbitrary percentiles of the values in the
-// sample.
-func (h *StandardHistogram) Percentiles(ps []float64) []float64 {
-	return h.sample.Percentiles(ps)
-}
-
-// Sample returns the Sample underlying the histogram.
-func (h *StandardHistogram) Sample() Sample { return h.sample }
-
 // Snapshot returns a read-only copy of the histogram.
-func (h *StandardHistogram) Snapshot() Histogram {
-	return &HistogramSnapshot{sample: h.sample.Snapshot().(*SampleSnapshot)}
+func (h *StandardHistogram) Snapshot() HistogramSnapshot {
+	return h.sample.Snapshot()
 }
 
-// StdDev returns the standard deviation of the values in the sample.
-func (h *StandardHistogram) StdDev() float64 { return h.sample.StdDev() }
-
-// Sum returns the sum in the sample.
-func (h *StandardHistogram) Sum() int64 { return h.sample.Sum() }
-
 // Update samples a new value.
 func (h *StandardHistogram) Update(v int64) { h.sample.Update(v) }
-
-// Variance returns the variance of the values in the sample.
-func (h *StandardHistogram) Variance() float64 { return h.sample.Variance() }
diff --git a/metrics/histogram_test.go b/metrics/histogram_test.go
index 7c9f42fcec..22fc5468b0 100644
--- a/metrics/histogram_test.go
+++ b/metrics/histogram_test.go
@@ -14,7 +14,7 @@ func TestGetOrRegisterHistogram(t *testing.T) {
 	r := NewRegistry()
 	s := NewUniformSample(100)
 	NewRegisteredHistogram("foo", r, s).Update(47)
-	if h := GetOrRegisterHistogram("foo", r, s); h.Count() != 1 {
+	if h := GetOrRegisterHistogram("foo", r, s).Snapshot(); h.Count() != 1 {
 		t.Fatal(h)
 	}
 }
@@ -24,11 +24,11 @@ func TestHistogram10000(t *testing.T) {
 	for i := 1; i <= 10000; i++ {
 		h.Update(int64(i))
 	}
-	testHistogram10000(t, h)
+	testHistogram10000(t, h.Snapshot())
 }
 
 func TestHistogramEmpty(t *testing.T) {
-	h := NewHistogram(NewUniformSample(100))
+	h := NewHistogram(NewUniformSample(100)).Snapshot()
 	if count := h.Count(); count != 0 {
 		t.Errorf("h.Count(): 0 != %v\n", count)
 	}
@@ -66,7 +66,7 @@ func TestHistogramSnapshot(t *testing.T) {
 	testHistogram10000(t, snapshot)
 }
 
-func testHistogram10000(t *testing.T, h Histogram) {
+func testHistogram10000(t *testing.T, h HistogramSnapshot) {
 	if count := h.Count(); count != 10000 {
 		t.Errorf("h.Count(): 10000 != %v\n", count)
 	}
diff --git a/metrics/inactive.go b/metrics/inactive.go
new file mode 100644
index 0000000000..1f47f0210a
--- /dev/null
+++ b/metrics/inactive.go
@@ -0,0 +1,48 @@
+// Copyright 2023 The go-ethereum Authors
+// This file is part of the go-ethereum library.
+//
+// The go-ethereum library is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Lesser General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// The go-ethereum library is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+// GNU Lesser General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public License
+// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
+
+package metrics
+
+// compile-time checks that interfaces are implemented.
+var (
+	_ SampleSnapshot    = (*emptySnapshot)(nil)
+	_ HistogramSnapshot = (*emptySnapshot)(nil)
+	_ CounterSnapshot   = (*emptySnapshot)(nil)
+	_ GaugeSnapshot     = (*emptySnapshot)(nil)
+	_ MeterSnapshot     = (*emptySnapshot)(nil)
+	_ EWMASnapshot      = (*emptySnapshot)(nil)
+	_ TimerSnapshot     = (*emptySnapshot)(nil)
+)
+
+type emptySnapshot struct{}
+
+func (*emptySnapshot) Count() int64                       { return 0 }
+func (*emptySnapshot) Max() int64                         { return 0 }
+func (*emptySnapshot) Mean() float64                      { return 0.0 }
+func (*emptySnapshot) Min() int64                         { return 0 }
+func (*emptySnapshot) Percentile(p float64) float64       { return 0.0 }
+func (*emptySnapshot) Percentiles(ps []float64) []float64 { return make([]float64, len(ps)) }
+func (*emptySnapshot) Size() int                          { return 0 }
+func (*emptySnapshot) StdDev() float64                    { return 0.0 }
+func (*emptySnapshot) Sum() int64                         { return 0 }
+func (*emptySnapshot) Values() []int64                    { return []int64{} }
+func (*emptySnapshot) Variance() float64                  { return 0.0 }
+func (*emptySnapshot) Value() int64                       { return 0 }
+func (*emptySnapshot) Rate() float64                      { return 0.0 }
+func (*emptySnapshot) Rate1() float64                     { return 0.0 }
+func (*emptySnapshot) Rate5() float64                     { return 0.0 }
+func (*emptySnapshot) Rate15() float64                    { return 0.0 }
+func (*emptySnapshot) RateMean() float64                  { return 0.0 }
diff --git a/metrics/influxdb/influxdb.go b/metrics/influxdb/influxdb.go
index 9354f1a633..bbc4fc024b 100644
--- a/metrics/influxdb/influxdb.go
+++ b/metrics/influxdb/influxdb.go
@@ -11,13 +11,13 @@ func readMeter(namespace, name string, i interface{}) (string, map[string]interf
 	case metrics.Counter:
 		measurement := fmt.Sprintf("%s%s.count", namespace, name)
 		fields := map[string]interface{}{
-			"value": metric.Count(),
+			"value": metric.Snapshot().Count(),
 		}
 		return measurement, fields
 	case metrics.CounterFloat64:
 		measurement := fmt.Sprintf("%s%s.count", namespace, name)
 		fields := map[string]interface{}{
-			"value": metric.Count(),
+			"value": metric.Snapshot().Count(),
 		}
 		return measurement, fields
 	case metrics.Gauge:
@@ -99,20 +99,19 @@ func readMeter(namespace, name string, i interface{}) (string, map[string]interf
 		return measurement, fields
 	case metrics.ResettingTimer:
 		t := metric.Snapshot()
-		if len(t.Values()) == 0 {
+		if t.Count() == 0 {
 			break
 		}
-		ps := t.Percentiles([]float64{50, 95, 99})
-		val := t.Values()
+		ps := t.Percentiles([]float64{0.50, 0.95, 0.99})
 		measurement := fmt.Sprintf("%s%s.span", namespace, name)
 		fields := map[string]interface{}{
-			"count": len(val),
-			"max":   val[len(val)-1],
+			"count": t.Count(),
+			"max":   t.Max(),
 			"mean":  t.Mean(),
-			"min":   val[0],
-			"p50":   ps[0],
-			"p95":   ps[1],
-			"p99":   ps[2],
+			"min":   t.Min(),
+			"p50":   int(ps[0]),
+			"p95":   int(ps[1]),
+			"p99":   int(ps[2]),
 		}
 		return measurement, fields
 	}
diff --git a/metrics/influxdb/influxdb_test.go b/metrics/influxdb/influxdb_test.go
index beeb36a531..c6f2eeac62 100644
--- a/metrics/influxdb/influxdb_test.go
+++ b/metrics/influxdb/influxdb_test.go
@@ -96,7 +96,7 @@ func TestExampleV2(t *testing.T) {
 	}
 	if have != want {
 		t.Errorf("\nhave:\n%v\nwant:\n%v\n", have, want)
-		t.Logf("have vs want:\n %v", findFirstDiffPos(have, want))
+		t.Logf("have vs want:\n%v", findFirstDiffPos(have, want))
 	}
 }
 
diff --git a/metrics/influxdb/testdata/influxdbv1.want b/metrics/influxdb/testdata/influxdbv1.want
index 5efffb9595..9443faedc5 100644
--- a/metrics/influxdb/testdata/influxdbv1.want
+++ b/metrics/influxdb/testdata/influxdbv1.want
@@ -1,3 +1,5 @@
+goth.system/cpu/schedlatency.histogram count=5645i,max=41943040i,mean=1819544.0410983171,min=0i,p25=0,p50=0,p75=7168,p95=16777216,p99=29360128,p999=33554432,p9999=33554432,stddev=6393570.217198883,variance=40877740122252.57 978307200000000000
+goth.system/memory/pauses.histogram count=14i,max=229376i,mean=50066.28571428572,min=5120i,p25=10240,p50=32768,p75=57344,p95=196608,p99=196608,p999=196608,p9999=196608,stddev=54726.062410783874,variance=2994941906.9890113 978307200000000000
 goth.test/counter.count value=12345 978307200000000000
 goth.test/counter_float64.count value=54321.98 978307200000000000
 goth.test/gauge.gauge value=23456i 978307200000000000
@@ -5,5 +7,5 @@ goth.test/gauge_float64.gauge value=34567.89 978307200000000000
 goth.test/gauge_info.gauge value="{\"arch\":\"amd64\",\"commit\":\"7caa2d8163ae3132c1c2d6978c76610caee2d949\",\"os\":\"linux\",\"protocol_versions\":\"64 65 66\",\"version\":\"1.10.18-unstable\"}" 978307200000000000
 goth.test/histogram.histogram count=3i,max=3i,mean=2,min=1i,p25=1,p50=2,p75=3,p95=3,p99=3,p999=3,p9999=3,stddev=0.816496580927726,variance=0.6666666666666666 978307200000000000
 goth.test/meter.meter count=0i,m1=0,m15=0,m5=0,mean=0 978307200000000000
-goth.test/resetting_timer.span count=6i,max=120000000i,mean=30000000,min=10000000i,p50=12000000i,p95=120000000i,p99=120000000i 978307200000000000
+goth.test/resetting_timer.span count=6i,max=120000000i,mean=30000000,min=10000000i,p50=12500000i,p95=120000000i,p99=120000000i 978307200000000000
 goth.test/timer.timer count=6i,m1=0,m15=0,m5=0,max=120000000i,mean=38333333.333333336,meanrate=0,min=20000000i,p50=22500000,p75=48000000,p95=120000000,p99=120000000,p999=120000000,p9999=120000000,stddev=36545253.529775314,variance=1335555555555555.2 978307200000000000
diff --git a/metrics/influxdb/testdata/influxdbv2.want b/metrics/influxdb/testdata/influxdbv2.want
index 5efffb9595..9443faedc5 100644
--- a/metrics/influxdb/testdata/influxdbv2.want
+++ b/metrics/influxdb/testdata/influxdbv2.want
@@ -1,3 +1,5 @@
+goth.system/cpu/schedlatency.histogram count=5645i,max=41943040i,mean=1819544.0410983171,min=0i,p25=0,p50=0,p75=7168,p95=16777216,p99=29360128,p999=33554432,p9999=33554432,stddev=6393570.217198883,variance=40877740122252.57 978307200000000000
+goth.system/memory/pauses.histogram count=14i,max=229376i,mean=50066.28571428572,min=5120i,p25=10240,p50=32768,p75=57344,p95=196608,p99=196608,p999=196608,p9999=196608,stddev=54726.062410783874,variance=2994941906.9890113 978307200000000000
 goth.test/counter.count value=12345 978307200000000000
 goth.test/counter_float64.count value=54321.98 978307200000000000
 goth.test/gauge.gauge value=23456i 978307200000000000
@@ -5,5 +7,5 @@ goth.test/gauge_float64.gauge value=34567.89 978307200000000000
 goth.test/gauge_info.gauge value="{\"arch\":\"amd64\",\"commit\":\"7caa2d8163ae3132c1c2d6978c76610caee2d949\",\"os\":\"linux\",\"protocol_versions\":\"64 65 66\",\"version\":\"1.10.18-unstable\"}" 978307200000000000
 goth.test/histogram.histogram count=3i,max=3i,mean=2,min=1i,p25=1,p50=2,p75=3,p95=3,p99=3,p999=3,p9999=3,stddev=0.816496580927726,variance=0.6666666666666666 978307200000000000
 goth.test/meter.meter count=0i,m1=0,m15=0,m5=0,mean=0 978307200000000000
-goth.test/resetting_timer.span count=6i,max=120000000i,mean=30000000,min=10000000i,p50=12000000i,p95=120000000i,p99=120000000i 978307200000000000
+goth.test/resetting_timer.span count=6i,max=120000000i,mean=30000000,min=10000000i,p50=12500000i,p95=120000000i,p99=120000000i 978307200000000000
 goth.test/timer.timer count=6i,m1=0,m15=0,m5=0,max=120000000i,mean=38333333.333333336,meanrate=0,min=20000000i,p50=22500000,p75=48000000,p95=120000000,p99=120000000,p999=120000000,p9999=120000000,stddev=36545253.529775314,variance=1335555555555555.2 978307200000000000
diff --git a/metrics/internal/sampledata.go b/metrics/internal/sampledata.go
index 9ace069576..de9b207b6d 100644
--- a/metrics/internal/sampledata.go
+++ b/metrics/internal/sampledata.go
@@ -17,6 +17,9 @@
 package internal
 
 import (
+	"bytes"
+	"encoding/gob"
+	metrics2 "runtime/metrics"
 	"time"
 
 	"github.com/ethereum/go-ethereum/metrics"
@@ -38,7 +41,15 @@ func ExampleMetrics() metrics.Registry {
 			"commit":            "7caa2d8163ae3132c1c2d6978c76610caee2d949",
 			"protocol_versions": "64 65 66",
 		})
-	metrics.NewRegisteredHistogram("test/histogram", registry, metrics.NewSampleSnapshot(3, []int64{1, 2, 3}))
+
+	{
+		s := metrics.NewUniformSample(3)
+		s.Update(1)
+		s.Update(2)
+		s.Update(3)
+		//metrics.NewRegisteredHistogram("test/histogram", registry, metrics.NewSampleSnapshot(3, []int64{1, 2, 3}))
+		metrics.NewRegisteredHistogram("test/histogram", registry, s)
+	}
 	registry.Register("test/meter", metrics.NewInactiveMeter())
 	{
 		timer := metrics.NewRegisteredResettingTimer("test/resetting_timer", registry)
@@ -60,5 +71,25 @@ func ExampleMetrics() metrics.Registry {
 		timer.Stop()
 	}
 	registry.Register("test/empty_resetting_timer", metrics.NewResettingTimer().Snapshot())
+
+	{ // go runtime metrics
+		var sLatency = "7\xff\x81\x03\x01\x01\x10Float64Histogram\x01\xff\x82\x00\x01\x02\x01\x06Counts\x01\xff\x84\x00\x01\aBuckets\x01\xff\x86\x00\x00\x00\x16\xff\x83\x02\x01\x01\b[]uint64\x01\xff\x84\x00\x01\x06\x00\x00\x17\xff\x85\x02\x01\x01\t[]float64\x01\xff\x86\x00\x01\b\x00\x00\xfe\x06T\xff\x82\x01\xff\xa2\x00\xfe\r\xef\x00\x01\x02\x02\x04\x05\x04\b\x15\x17 B?6.L;$!2) \x1a? \x190aH7FY6#\x190\x1d\x14\x10\x1b\r\t\x04\x03\x01\x01\x00\x03\x02\x00\x03\x05\x05\x02\x02\x06\x04\v\x06\n\x15\x18\x13'&.\x12=H/L&\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\xff\xa3\xfe\xf0\xff\x00\xf8\x95\xd6&\xe8\v.q>\xf8\x95\xd6&\xe8\v.\x81>\xf8\xdfA:\xdc\x11ʼn>\xf8\x95\xd6&\xe8\v.\x91>\xf8:\x8c0\xe2\x8ey\x95>\xf8\xdfA:\xdc\x11ř>\xf8\x84\xf7C֔\x10\x9e>\xf8\x95\xd6&\xe8\v.\xa1>\xf8:\x8c0\xe2\x8ey\xa5>\xf8\xdfA:\xdc\x11ũ>\xf8\x84\xf7C֔\x10\xae>\xf8\x95\xd6&\xe8\v.\xb1>\xf8:\x8c0\xe2\x8ey\xb5>\xf8\xdfA:\xdc\x11Ź>\xf8\x84\xf7C֔\x10\xbe>\xf8\x95\xd6&\xe8\v.\xc1>\xf8:\x8c0\xe2\x8ey\xc5>\xf8\xdfA:\xdc\x11\xc5\xc9>\xf8\x84\xf7C֔\x10\xce>\xf8\x95\xd6&\xe8\v.\xd1>\xf8:\x8c0\xe2\x8ey\xd5>\xf8\xdfA:\xdc\x11\xc5\xd9>\xf8\x84\xf7C֔\x10\xde>\xf8\x95\xd6&\xe8\v.\xe1>\xf8:\x8c0\xe2\x8ey\xe5>\xf8\xdfA:\xdc\x11\xc5\xe9>\xf8\x84\xf7C֔\x10\xee>\xf8\x95\xd6&\xe8\v.\xf1>\xf8:\x8c0\xe2\x8ey\xf5>\xf8\xdfA:\xdc\x11\xc5\xf9>\xf8\x84\xf7C֔\x10\xfe>\xf8\x95\xd6&\xe8\v.\x01?\xf8:\x8c0\xe2\x8ey\x05?\xf8\xdfA:\xdc\x11\xc5\t?\xf8\x84\xf7C֔\x10\x0e?\xf8\x95\xd6&\xe8\v.\x11?\xf8:\x8c0\xe2\x8ey\x15?\xf8\xdfA:\xdc\x11\xc5\x19?\xf8\x84\xf7C֔\x10\x1e?\xf8\x95\xd6&\xe8\v.!?\xf8:\x8c0\xe2\x8ey%?\xf8\xdfA:\xdc\x11\xc5)?\xf8\x84\xf7C֔\x10.?\xf8\x95\xd6&\xe8\v.1?\xf8:\x8c0\xe2\x8ey5?\xf8\xdfA:\xdc\x11\xc59?\xf8\x84\xf7C֔\x10>?\xf8\x95\xd6&\xe8\v.A?\xf8:\x8c0\xe2\x8eyE?\xf8\xdfA:\xdc\x11\xc5I?\xf8\x84\xf7C֔\x10N?\xf8\x95\xd6&\xe8\v.Q?\xf8:\x8c0\xe2\x8eyU?\xf8\xdfA:\xdc\x11\xc5Y?\xf8\x84\xf7C֔\x10^?\xf8\x95\xd6&\xe8\v.a?\xf8:\x8c0\xe2\x8eye?\xf8\xdfA:\xdc\x11\xc5i?\xf8\x84\xf7C֔\x10n?\xf8\x95\xd6&\xe8\v.q?\xf8:\x8c0\xe2\x8eyu?\xf8\xdfA:\xdc\x11\xc5y?\xf8\x84\xf7C֔\x10~?\xf8\x95\xd6&\xe8\v.\x81?\xf8:\x8c0\xe2\x8ey\x85?\xf8\xdfA:\xdc\x11ʼn?\xf8\x84\xf7C֔\x10\x8e?\xf8\x95\xd6&\xe8\v.\x91?\xf8:\x8c0\xe2\x8ey\x95?\xf8\xdfA:\xdc\x11ř?\xf8\x84\xf7C֔\x10\x9e?\xf8\x95\xd6&\xe8\v.\xa1?\xf8:\x8c0\xe2\x8ey\xa5?\xf8\xdfA:\xdc\x11ũ?\xf8\x84\xf7C֔\x10\xae?\xf8\x95\xd6&\xe8\v.\xb1?\xf8:\x8c0\xe2\x8ey\xb5?\xf8\xdfA:\xdc\x11Ź?\xf8\x84\xf7C֔\x10\xbe?\xf8\x95\xd6&\xe8\v.\xc1?\xf8:\x8c0\xe2\x8ey\xc5?\xf8\xdfA:\xdc\x11\xc5\xc9?\xf8\x84\xf7C֔\x10\xce?\xf8\x95\xd6&\xe8\v.\xd1?\xf8:\x8c0\xe2\x8ey\xd5?\xf8\xdfA:\xdc\x11\xc5\xd9?\xf8\x84\xf7C֔\x10\xde?\xf8\x95\xd6&\xe8\v.\xe1?\xf8:\x8c0\xe2\x8ey\xe5?\xf8\xdfA:\xdc\x11\xc5\xe9?\xf8\x84\xf7C֔\x10\xee?\xf8\x95\xd6&\xe8\v.\xf1?\xf8:\x8c0\xe2\x8ey\xf5?\xf8\xdfA:\xdc\x11\xc5\xf9?\xf8\x84\xf7C֔\x10\xfe?\xf8\x95\xd6&\xe8\v.\x01@\xf8:\x8c0\xe2\x8ey\x05@\xf8\xdfA:\xdc\x11\xc5\t@\xf8\x84\xf7C֔\x10\x0e@\xf8\x95\xd6&\xe8\v.\x11@\xf8:\x8c0\xe2\x8ey\x15@\xf8\xdfA:\xdc\x11\xc5\x19@\xf8\x84\xf7C֔\x10\x1e@\xf8\x95\xd6&\xe8\v.!@\xf8:\x8c0\xe2\x8ey%@\xf8\xdfA:\xdc\x11\xc5)@\xf8\x84\xf7C֔\x10.@\xf8\x95\xd6&\xe8\v.1@\xf8:\x8c0\xe2\x8ey5@\xf8\xdfA:\xdc\x11\xc59@\xf8\x84\xf7C֔\x10>@\xf8\x95\xd6&\xe8\v.A@\xf8:\x8c0\xe2\x8eyE@\xf8\xdfA:\xdc\x11\xc5I@\xf8\x84\xf7C֔\x10N@\xf8\x95\xd6&\xe8\v.Q@\xf8:\x8c0\xe2\x8eyU@\xf8\xdfA:\xdc\x11\xc5Y@\xf8\x84\xf7C֔\x10^@\xf8\x95\xd6&\xe8\v.a@\xf8:\x8c0\xe2\x8eye@\xf8\xdfA:\xdc\x11\xc5i@\xf8\x84\xf7C֔\x10n@\xf8\x95\xd6&\xe8\v.q@\xf8:\x8c0\xe2\x8eyu@\xf8\xdfA:\xdc\x11\xc5y@\xf8\x84\xf7C֔\x10~@\xf8\x95\xd6&\xe8\v.\x81@\xf8:\x8c0\xe2\x8ey\x85@\xf8\xdfA:\xdc\x11ʼn@\xf8\x84\xf7C֔\x10\x8e@\xf8\x95\xd6&\xe8\v.\x91@\xf8:\x8c0\xe2\x8ey\x95@\xf8\xdfA:\xdc\x11ř@\xf8\x84\xf7C֔\x10\x9e@\xf8\x95\xd6&\xe8\v.\xa1@\xf8:\x8c0\xe2\x8ey\xa5@\xf8\xdfA:\xdc\x11ũ@\xf8\x84\xf7C֔\x10\xae@\xf8\x95\xd6&\xe8\v.\xb1@\xf8:\x8c0\xe2\x8ey\xb5@\xf8\xdfA:\xdc\x11Ź@\xf8\x84\xf7C֔\x10\xbe@\xf8\x95\xd6&\xe8\v.\xc1@\xf8:\x8c0\xe2\x8ey\xc5@\xf8\xdfA:\xdc\x11\xc5\xc9@\xf8\x84\xf7C֔\x10\xce@\xf8\x95\xd6&\xe8\v.\xd1@\xf8:\x8c0\xe2\x8ey\xd5@\xf8\xdfA:\xdc\x11\xc5\xd9@\xf8\x84\xf7C֔\x10\xde@\xf8\x95\xd6&\xe8\v.\xe1@\xf8:\x8c0\xe2\x8ey\xe5@\xf8\xdfA:\xdc\x11\xc5\xe9@\xf8\x84\xf7C֔\x10\xee@\xf8\x95\xd6&\xe8\v.\xf1@\xf8:\x8c0\xe2\x8ey\xf5@\xf8\xdfA:\xdc\x11\xc5\xf9@\xf8\x84\xf7C֔\x10\xfe@\xf8\x95\xd6&\xe8\v.\x01A\xfe\xf0\x7f\x00"
+		var gcPauses = "7\xff\x81\x03\x01\x01\x10Float64Histogram\x01\xff\x82\x00\x01\x02\x01\x06Counts\x01\xff\x84\x00\x01\aBuckets\x01\xff\x86\x00\x00\x00\x16\xff\x83\x02\x01\x01\b[]uint64\x01\xff\x84\x00\x01\x06\x00\x00\x17\xff\x85\x02\x01\x01\t[]float64\x01\xff\x86\x00\x01\b\x00\x00\xfe\x06R\xff\x82\x01\xff\xa2\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\x01\x01\x00\x01\x00\x00\x00\x01\x01\x01\x01\x01\x01\x01\x00\x02\x00\x00\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\xff\xa3\xfe\xf0\xff\x00\xf8\x95\xd6&\xe8\v.q>\xf8\x95\xd6&\xe8\v.\x81>\xf8\xdfA:\xdc\x11ʼn>\xf8\x95\xd6&\xe8\v.\x91>\xf8:\x8c0\xe2\x8ey\x95>\xf8\xdfA:\xdc\x11ř>\xf8\x84\xf7C֔\x10\x9e>\xf8\x95\xd6&\xe8\v.\xa1>\xf8:\x8c0\xe2\x8ey\xa5>\xf8\xdfA:\xdc\x11ũ>\xf8\x84\xf7C֔\x10\xae>\xf8\x95\xd6&\xe8\v.\xb1>\xf8:\x8c0\xe2\x8ey\xb5>\xf8\xdfA:\xdc\x11Ź>\xf8\x84\xf7C֔\x10\xbe>\xf8\x95\xd6&\xe8\v.\xc1>\xf8:\x8c0\xe2\x8ey\xc5>\xf8\xdfA:\xdc\x11\xc5\xc9>\xf8\x84\xf7C֔\x10\xce>\xf8\x95\xd6&\xe8\v.\xd1>\xf8:\x8c0\xe2\x8ey\xd5>\xf8\xdfA:\xdc\x11\xc5\xd9>\xf8\x84\xf7C֔\x10\xde>\xf8\x95\xd6&\xe8\v.\xe1>\xf8:\x8c0\xe2\x8ey\xe5>\xf8\xdfA:\xdc\x11\xc5\xe9>\xf8\x84\xf7C֔\x10\xee>\xf8\x95\xd6&\xe8\v.\xf1>\xf8:\x8c0\xe2\x8ey\xf5>\xf8\xdfA:\xdc\x11\xc5\xf9>\xf8\x84\xf7C֔\x10\xfe>\xf8\x95\xd6&\xe8\v.\x01?\xf8:\x8c0\xe2\x8ey\x05?\xf8\xdfA:\xdc\x11\xc5\t?\xf8\x84\xf7C֔\x10\x0e?\xf8\x95\xd6&\xe8\v.\x11?\xf8:\x8c0\xe2\x8ey\x15?\xf8\xdfA:\xdc\x11\xc5\x19?\xf8\x84\xf7C֔\x10\x1e?\xf8\x95\xd6&\xe8\v.!?\xf8:\x8c0\xe2\x8ey%?\xf8\xdfA:\xdc\x11\xc5)?\xf8\x84\xf7C֔\x10.?\xf8\x95\xd6&\xe8\v.1?\xf8:\x8c0\xe2\x8ey5?\xf8\xdfA:\xdc\x11\xc59?\xf8\x84\xf7C֔\x10>?\xf8\x95\xd6&\xe8\v.A?\xf8:\x8c0\xe2\x8eyE?\xf8\xdfA:\xdc\x11\xc5I?\xf8\x84\xf7C֔\x10N?\xf8\x95\xd6&\xe8\v.Q?\xf8:\x8c0\xe2\x8eyU?\xf8\xdfA:\xdc\x11\xc5Y?\xf8\x84\xf7C֔\x10^?\xf8\x95\xd6&\xe8\v.a?\xf8:\x8c0\xe2\x8eye?\xf8\xdfA:\xdc\x11\xc5i?\xf8\x84\xf7C֔\x10n?\xf8\x95\xd6&\xe8\v.q?\xf8:\x8c0\xe2\x8eyu?\xf8\xdfA:\xdc\x11\xc5y?\xf8\x84\xf7C֔\x10~?\xf8\x95\xd6&\xe8\v.\x81?\xf8:\x8c0\xe2\x8ey\x85?\xf8\xdfA:\xdc\x11ʼn?\xf8\x84\xf7C֔\x10\x8e?\xf8\x95\xd6&\xe8\v.\x91?\xf8:\x8c0\xe2\x8ey\x95?\xf8\xdfA:\xdc\x11ř?\xf8\x84\xf7C֔\x10\x9e?\xf8\x95\xd6&\xe8\v.\xa1?\xf8:\x8c0\xe2\x8ey\xa5?\xf8\xdfA:\xdc\x11ũ?\xf8\x84\xf7C֔\x10\xae?\xf8\x95\xd6&\xe8\v.\xb1?\xf8:\x8c0\xe2\x8ey\xb5?\xf8\xdfA:\xdc\x11Ź?\xf8\x84\xf7C֔\x10\xbe?\xf8\x95\xd6&\xe8\v.\xc1?\xf8:\x8c0\xe2\x8ey\xc5?\xf8\xdfA:\xdc\x11\xc5\xc9?\xf8\x84\xf7C֔\x10\xce?\xf8\x95\xd6&\xe8\v.\xd1?\xf8:\x8c0\xe2\x8ey\xd5?\xf8\xdfA:\xdc\x11\xc5\xd9?\xf8\x84\xf7C֔\x10\xde?\xf8\x95\xd6&\xe8\v.\xe1?\xf8:\x8c0\xe2\x8ey\xe5?\xf8\xdfA:\xdc\x11\xc5\xe9?\xf8\x84\xf7C֔\x10\xee?\xf8\x95\xd6&\xe8\v.\xf1?\xf8:\x8c0\xe2\x8ey\xf5?\xf8\xdfA:\xdc\x11\xc5\xf9?\xf8\x84\xf7C֔\x10\xfe?\xf8\x95\xd6&\xe8\v.\x01@\xf8:\x8c0\xe2\x8ey\x05@\xf8\xdfA:\xdc\x11\xc5\t@\xf8\x84\xf7C֔\x10\x0e@\xf8\x95\xd6&\xe8\v.\x11@\xf8:\x8c0\xe2\x8ey\x15@\xf8\xdfA:\xdc\x11\xc5\x19@\xf8\x84\xf7C֔\x10\x1e@\xf8\x95\xd6&\xe8\v.!@\xf8:\x8c0\xe2\x8ey%@\xf8\xdfA:\xdc\x11\xc5)@\xf8\x84\xf7C֔\x10.@\xf8\x95\xd6&\xe8\v.1@\xf8:\x8c0\xe2\x8ey5@\xf8\xdfA:\xdc\x11\xc59@\xf8\x84\xf7C֔\x10>@\xf8\x95\xd6&\xe8\v.A@\xf8:\x8c0\xe2\x8eyE@\xf8\xdfA:\xdc\x11\xc5I@\xf8\x84\xf7C֔\x10N@\xf8\x95\xd6&\xe8\v.Q@\xf8:\x8c0\xe2\x8eyU@\xf8\xdfA:\xdc\x11\xc5Y@\xf8\x84\xf7C֔\x10^@\xf8\x95\xd6&\xe8\v.a@\xf8:\x8c0\xe2\x8eye@\xf8\xdfA:\xdc\x11\xc5i@\xf8\x84\xf7C֔\x10n@\xf8\x95\xd6&\xe8\v.q@\xf8:\x8c0\xe2\x8eyu@\xf8\xdfA:\xdc\x11\xc5y@\xf8\x84\xf7C֔\x10~@\xf8\x95\xd6&\xe8\v.\x81@\xf8:\x8c0\xe2\x8ey\x85@\xf8\xdfA:\xdc\x11ʼn@\xf8\x84\xf7C֔\x10\x8e@\xf8\x95\xd6&\xe8\v.\x91@\xf8:\x8c0\xe2\x8ey\x95@\xf8\xdfA:\xdc\x11ř@\xf8\x84\xf7C֔\x10\x9e@\xf8\x95\xd6&\xe8\v.\xa1@\xf8:\x8c0\xe2\x8ey\xa5@\xf8\xdfA:\xdc\x11ũ@\xf8\x84\xf7C֔\x10\xae@\xf8\x95\xd6&\xe8\v.\xb1@\xf8:\x8c0\xe2\x8ey\xb5@\xf8\xdfA:\xdc\x11Ź@\xf8\x84\xf7C֔\x10\xbe@\xf8\x95\xd6&\xe8\v.\xc1@\xf8:\x8c0\xe2\x8ey\xc5@\xf8\xdfA:\xdc\x11\xc5\xc9@\xf8\x84\xf7C֔\x10\xce@\xf8\x95\xd6&\xe8\v.\xd1@\xf8:\x8c0\xe2\x8ey\xd5@\xf8\xdfA:\xdc\x11\xc5\xd9@\xf8\x84\xf7C֔\x10\xde@\xf8\x95\xd6&\xe8\v.\xe1@\xf8:\x8c0\xe2\x8ey\xe5@\xf8\xdfA:\xdc\x11\xc5\xe9@\xf8\x84\xf7C֔\x10\xee@\xf8\x95\xd6&\xe8\v.\xf1@\xf8:\x8c0\xe2\x8ey\xf5@\xf8\xdfA:\xdc\x11\xc5\xf9@\xf8\x84\xf7C֔\x10\xfe@\xf8\x95\xd6&\xe8\v.\x01A\xfe\xf0\x7f\x00"
+
+		var secondsToNs = float64(time.Second)
+
+		dserialize := func(data string) *metrics2.Float64Histogram {
+			var res metrics2.Float64Histogram
+			if err := gob.NewDecoder(bytes.NewReader([]byte(data))).Decode(&res); err != nil {
+				panic(err)
+			}
+			return &res
+		}
+		cpuSchedLatency := metrics.RuntimeHistogramFromData(secondsToNs, dserialize(sLatency))
+		registry.Register("system/cpu/schedlatency", cpuSchedLatency)
+
+		memPauses := metrics.RuntimeHistogramFromData(secondsToNs, dserialize(gcPauses))
+		registry.Register("system/memory/pauses", memPauses)
+	}
 	return registry
 }
diff --git a/metrics/internal/sampledata_test.go b/metrics/internal/sampledata_test.go
new file mode 100644
index 0000000000..0013299406
--- /dev/null
+++ b/metrics/internal/sampledata_test.go
@@ -0,0 +1,27 @@
+package internal
+
+import (
+	"bytes"
+	"encoding/gob"
+	"fmt"
+	metrics2 "runtime/metrics"
+	"testing"
+	"time"
+
+	"github.com/ethereum/go-ethereum/metrics"
+)
+
+func TestCollectRuntimeMetrics(t *testing.T) {
+	t.Skip("Only used for generating testdata")
+	serialize := func(path string, histogram *metrics2.Float64Histogram) {
+		var f = new(bytes.Buffer)
+		if err := gob.NewEncoder(f).Encode(histogram); err != nil {
+			panic(err)
+		}
+		fmt.Printf("var %v = %q\n", path, f.Bytes())
+	}
+	time.Sleep(2 * time.Second)
+	stats := metrics.ReadRuntimeStats()
+	serialize("schedlatency", stats.SchedLatency)
+	serialize("gcpauses", stats.GCPauses)
+}
diff --git a/metrics/librato/librato.go b/metrics/librato/librato.go
index fa98595991..a86f758637 100644
--- a/metrics/librato/librato.go
+++ b/metrics/librato/librato.go
@@ -61,16 +61,16 @@ func (rep *Reporter) Run() {
 
 // calculate sum of squares from data provided by metrics.Histogram
 // see http://en.wikipedia.org/wiki/Standard_deviation#Rapid_calculation_methods
-func sumSquares(s metrics.Sample) float64 {
-	count := float64(s.Count())
-	sumSquared := math.Pow(count*s.Mean(), 2)
-	sumSquares := math.Pow(count*s.StdDev(), 2) + sumSquared/count
+func sumSquares(icount int64, mean, stDev float64) float64 {
+	count := float64(icount)
+	sumSquared := math.Pow(count*mean, 2)
+	sumSquares := math.Pow(count*stDev, 2) + sumSquared/count
 	if math.IsNaN(sumSquares) {
 		return 0.0
 	}
 	return sumSquares
 }
-func sumSquaresTimer(t metrics.Timer) float64 {
+func sumSquaresTimer(t metrics.TimerSnapshot) float64 {
 	count := float64(t.Count())
 	sumSquared := math.Pow(count*t.Mean(), 2)
 	sumSquares := math.Pow(count*t.StdDev(), 2) + sumSquared/count
@@ -97,9 +97,10 @@ func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot B
 		measurement[Period] = rep.Interval.Seconds()
 		switch m := metric.(type) {
 		case metrics.Counter:
-			if m.Count() > 0 {
+			ms := m.Snapshot()
+			if ms.Count() > 0 {
 				measurement[Name] = fmt.Sprintf("%s.%s", name, "count")
-				measurement[Value] = float64(m.Count())
+				measurement[Value] = float64(ms.Count())
 				measurement[Attributes] = map[string]interface{}{
 					DisplayUnitsLong:  Operations,
 					DisplayUnitsShort: OperationsShort,
@@ -108,9 +109,9 @@ func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot B
 				snapshot.Counters = append(snapshot.Counters, measurement)
 			}
 		case metrics.CounterFloat64:
-			if m.Count() > 0 {
+			if count := m.Snapshot().Count(); count > 0 {
 				measurement[Name] = fmt.Sprintf("%s.%s", name, "count")
-				measurement[Value] = m.Count()
+				measurement[Value] = count
 				measurement[Attributes] = map[string]interface{}{
 					DisplayUnitsLong:  Operations,
 					DisplayUnitsShort: OperationsShort,
@@ -120,44 +121,45 @@ func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot B
 			}
 		case metrics.Gauge:
 			measurement[Name] = name
-			measurement[Value] = float64(m.Value())
+			measurement[Value] = float64(m.Snapshot().Value())
 			snapshot.Gauges = append(snapshot.Gauges, measurement)
 		case metrics.GaugeFloat64:
 			measurement[Name] = name
-			measurement[Value] = m.Value()
+			measurement[Value] = m.Snapshot().Value()
 			snapshot.Gauges = append(snapshot.Gauges, measurement)
 		case metrics.GaugeInfo:
 			measurement[Name] = name
-			measurement[Value] = m.Value()
+			measurement[Value] = m.Snapshot().Value()
 			snapshot.Gauges = append(snapshot.Gauges, measurement)
 		case metrics.Histogram:
-			if m.Count() > 0 {
+			ms := m.Snapshot()
+			if ms.Count() > 0 {
 				gauges := make([]Measurement, histogramGaugeCount)
-				s := m.Sample()
 				measurement[Name] = fmt.Sprintf("%s.%s", name, "hist")
-				measurement[Count] = uint64(s.Count())
-				measurement[Max] = float64(s.Max())
-				measurement[Min] = float64(s.Min())
-				measurement[Sum] = float64(s.Sum())
-				measurement[SumSquares] = sumSquares(s)
+				measurement[Count] = uint64(ms.Count())
+				measurement[Max] = float64(ms.Max())
+				measurement[Min] = float64(ms.Min())
+				measurement[Sum] = float64(ms.Sum())
+				measurement[SumSquares] = sumSquares(ms.Count(), ms.Mean(), ms.StdDev())
 				gauges[0] = measurement
 				for i, p := range rep.Percentiles {
 					gauges[i+1] = Measurement{
 						Name:   fmt.Sprintf("%s.%.2f", measurement[Name], p),
-						Value:  s.Percentile(p),
+						Value:  ms.Percentile(p),
 						Period: measurement[Period],
 					}
 				}
 				snapshot.Gauges = append(snapshot.Gauges, gauges...)
 			}
 		case metrics.Meter:
+			ms := m.Snapshot()
 			measurement[Name] = name
-			measurement[Value] = float64(m.Count())
+			measurement[Value] = float64(ms.Count())
 			snapshot.Counters = append(snapshot.Counters, measurement)
 			snapshot.Gauges = append(snapshot.Gauges,
 				Measurement{
 					Name:   fmt.Sprintf("%s.%s", name, "1min"),
-					Value:  m.Rate1(),
+					Value:  ms.Rate1(),
 					Period: int64(rep.Interval.Seconds()),
 					Attributes: map[string]interface{}{
 						DisplayUnitsLong:  Operations,
@@ -167,7 +169,7 @@ func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot B
 				},
 				Measurement{
 					Name:   fmt.Sprintf("%s.%s", name, "5min"),
-					Value:  m.Rate5(),
+					Value:  ms.Rate5(),
 					Period: int64(rep.Interval.Seconds()),
 					Attributes: map[string]interface{}{
 						DisplayUnitsLong:  Operations,
@@ -177,7 +179,7 @@ func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot B
 				},
 				Measurement{
 					Name:   fmt.Sprintf("%s.%s", name, "15min"),
-					Value:  m.Rate15(),
+					Value:  ms.Rate15(),
 					Period: int64(rep.Interval.Seconds()),
 					Attributes: map[string]interface{}{
 						DisplayUnitsLong:  Operations,
@@ -187,26 +189,27 @@ func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot B
 				},
 			)
 		case metrics.Timer:
+			ms := m.Snapshot()
 			measurement[Name] = name
-			measurement[Value] = float64(m.Count())
+			measurement[Value] = float64(ms.Count())
 			snapshot.Counters = append(snapshot.Counters, measurement)
-			if m.Count() > 0 {
+			if ms.Count() > 0 {
 				libratoName := fmt.Sprintf("%s.%s", name, "timer.mean")
 				gauges := make([]Measurement, histogramGaugeCount)
 				gauges[0] = Measurement{
 					Name:       libratoName,
-					Count:      uint64(m.Count()),
-					Sum:        m.Mean() * float64(m.Count()),
-					Max:        float64(m.Max()),
-					Min:        float64(m.Min()),
-					SumSquares: sumSquaresTimer(m),
+					Count:      uint64(ms.Count()),
+					Sum:        ms.Mean() * float64(ms.Count()),
+					Max:        float64(ms.Max()),
+					Min:        float64(ms.Min()),
+					SumSquares: sumSquaresTimer(ms),
 					Period:     int64(rep.Interval.Seconds()),
 					Attributes: rep.TimerAttributes,
 				}
 				for i, p := range rep.Percentiles {
 					gauges[i+1] = Measurement{
 						Name:       fmt.Sprintf("%s.timer.%2.0f", name, p*100),
-						Value:      m.Percentile(p),
+						Value:      ms.Percentile(p),
 						Period:     int64(rep.Interval.Seconds()),
 						Attributes: rep.TimerAttributes,
 					}
@@ -215,7 +218,7 @@ func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot B
 				snapshot.Gauges = append(snapshot.Gauges,
 					Measurement{
 						Name:   fmt.Sprintf("%s.%s", name, "rate.1min"),
-						Value:  m.Rate1(),
+						Value:  ms.Rate1(),
 						Period: int64(rep.Interval.Seconds()),
 						Attributes: map[string]interface{}{
 							DisplayUnitsLong:  Operations,
@@ -225,7 +228,7 @@ func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot B
 					},
 					Measurement{
 						Name:   fmt.Sprintf("%s.%s", name, "rate.5min"),
-						Value:  m.Rate5(),
+						Value:  ms.Rate5(),
 						Period: int64(rep.Interval.Seconds()),
 						Attributes: map[string]interface{}{
 							DisplayUnitsLong:  Operations,
@@ -235,7 +238,7 @@ func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot B
 					},
 					Measurement{
 						Name:   fmt.Sprintf("%s.%s", name, "rate.15min"),
-						Value:  m.Rate15(),
+						Value:  ms.Rate15(),
 						Period: int64(rep.Interval.Seconds()),
 						Attributes: map[string]interface{}{
 							DisplayUnitsLong:  Operations,
diff --git a/metrics/log.go b/metrics/log.go
index d71a1c3d9c..3b9773faa7 100644
--- a/metrics/log.go
+++ b/metrics/log.go
@@ -23,19 +23,19 @@ func LogScaled(r Registry, freq time.Duration, scale time.Duration, l Logger) {
 			switch metric := i.(type) {
 			case Counter:
 				l.Printf("counter %s\n", name)
-				l.Printf("  count:       %9d\n", metric.Count())
+				l.Printf("  count:       %9d\n", metric.Snapshot().Count())
 			case CounterFloat64:
 				l.Printf("counter %s\n", name)
-				l.Printf("  count:       %f\n", metric.Count())
+				l.Printf("  count:       %f\n", metric.Snapshot().Count())
 			case Gauge:
 				l.Printf("gauge %s\n", name)
-				l.Printf("  value:       %9d\n", metric.Value())
+				l.Printf("  value:       %9d\n", metric.Snapshot().Value())
 			case GaugeFloat64:
 				l.Printf("gauge %s\n", name)
-				l.Printf("  value:       %f\n", metric.Value())
+				l.Printf("  value:       %f\n", metric.Snapshot().Value())
 			case GaugeInfo:
 				l.Printf("gauge %s\n", name)
-				l.Printf("  value:       %s\n", metric.Value())
+				l.Printf("  value:       %s\n", metric.Snapshot().Value())
 			case Healthcheck:
 				metric.Check()
 				l.Printf("healthcheck %s\n", name)
diff --git a/metrics/meter.go b/metrics/meter.go
index 8a89dc4275..22475ef6eb 100644
--- a/metrics/meter.go
+++ b/metrics/meter.go
@@ -1,21 +1,25 @@
 package metrics
 
 import (
+	"math"
 	"sync"
 	"sync/atomic"
 	"time"
 )
 
-// Meters count events to produce exponentially-weighted moving average rates
-// at one-, five-, and fifteen-minutes and a mean rate.
-type Meter interface {
+type MeterSnapshot interface {
 	Count() int64
-	Mark(int64)
 	Rate1() float64
 	Rate5() float64
 	Rate15() float64
 	RateMean() float64
-	Snapshot() Meter
+}
+
+// Meters count events to produce exponentially-weighted moving average rates
+// at one-, five-, and fifteen-minutes and a mean rate.
+type Meter interface {
+	Mark(int64)
+	Snapshot() MeterSnapshot
 	Stop()
 }
 
@@ -30,17 +34,6 @@ func GetOrRegisterMeter(name string, r Registry) Meter {
 	return r.GetOrRegister(name, NewMeter).(Meter)
 }
 
-// GetOrRegisterMeterForced returns an existing Meter or constructs and registers a
-// new StandardMeter no matter the global switch is enabled or not.
-// Be sure to unregister the meter from the registry once it is of no use to
-// allow for garbage collection.
-func GetOrRegisterMeterForced(name string, r Registry) Meter {
-	if nil == r {
-		r = DefaultRegistry
-	}
-	return r.GetOrRegister(name, NewMeterForced).(Meter)
-}
-
 // NewMeter constructs a new StandardMeter and launches a goroutine.
 // Be sure to call Stop() once the meter is of no use to allow for garbage collection.
 func NewMeter() Meter {
@@ -68,115 +61,53 @@ func NewInactiveMeter() Meter {
 	return m
 }
 
-// NewMeterForced constructs a new StandardMeter and launches a goroutine no matter
-// the global switch is enabled or not.
-// Be sure to call Stop() once the meter is of no use to allow for garbage collection.
-func NewMeterForced() Meter {
-	m := newStandardMeter()
-	arbiter.Lock()
-	defer arbiter.Unlock()
-	arbiter.meters[m] = struct{}{}
-	if !arbiter.started {
-		arbiter.started = true
-		go arbiter.tick()
-	}
-	return m
-}
-
 // NewRegisteredMeter constructs and registers a new StandardMeter
 // and launches a goroutine.
 // Be sure to unregister the meter from the registry once it is of no use to
 // allow for garbage collection.
 func NewRegisteredMeter(name string, r Registry) Meter {
-	c := NewMeter()
-	if nil == r {
-		r = DefaultRegistry
-	}
-	r.Register(name, c)
-	return c
+	return GetOrRegisterMeter(name, r)
 }
 
-// NewRegisteredMeterForced constructs and registers a new StandardMeter
-// and launches a goroutine no matter the global switch is enabled or not.
-// Be sure to unregister the meter from the registry once it is of no use to
-// allow for garbage collection.
-func NewRegisteredMeterForced(name string, r Registry) Meter {
-	c := NewMeterForced()
-	if nil == r {
-		r = DefaultRegistry
-	}
-	r.Register(name, c)
-	return c
-}
-
-// MeterSnapshot is a read-only copy of another Meter.
-type MeterSnapshot struct {
-	temp                           atomic.Int64
+// meterSnapshot is a read-only copy of the meter's internal values.
+type meterSnapshot struct {
 	count                          int64
 	rate1, rate5, rate15, rateMean float64
 }
 
 // Count returns the count of events at the time the snapshot was taken.
-func (m *MeterSnapshot) Count() int64 { return m.count }
-
-// Mark panics.
-func (*MeterSnapshot) Mark(n int64) {
-	panic("Mark called on a MeterSnapshot")
-}
+func (m *meterSnapshot) Count() int64 { return m.count }
 
 // Rate1 returns the one-minute moving average rate of events per second at the
 // time the snapshot was taken.
-func (m *MeterSnapshot) Rate1() float64 { return m.rate1 }
+func (m *meterSnapshot) Rate1() float64 { return m.rate1 }
 
 // Rate5 returns the five-minute moving average rate of events per second at
 // the time the snapshot was taken.
-func (m *MeterSnapshot) Rate5() float64 { return m.rate5 }
+func (m *meterSnapshot) Rate5() float64 { return m.rate5 }
 
 // Rate15 returns the fifteen-minute moving average rate of events per second
 // at the time the snapshot was taken.
-func (m *MeterSnapshot) Rate15() float64 { return m.rate15 }
+func (m *meterSnapshot) Rate15() float64 { return m.rate15 }
 
 // RateMean returns the meter's mean rate of events per second at the time the
 // snapshot was taken.
-func (m *MeterSnapshot) RateMean() float64 { return m.rateMean }
-
-// Snapshot returns the snapshot.
-func (m *MeterSnapshot) Snapshot() Meter { return m }
-
-// Stop is a no-op.
-func (m *MeterSnapshot) Stop() {}
+func (m *meterSnapshot) RateMean() float64 { return m.rateMean }
 
 // NilMeter is a no-op Meter.
 type NilMeter struct{}
 
-// Count is a no-op.
-func (NilMeter) Count() int64 { return 0 }
-
-// Mark is a no-op.
-func (NilMeter) Mark(n int64) {}
-
-// Rate1 is a no-op.
-func (NilMeter) Rate1() float64 { return 0.0 }
-
-// Rate5 is a no-op.
-func (NilMeter) Rate5() float64 { return 0.0 }
-
-// Rate15 is a no-op.
-func (NilMeter) Rate15() float64 { return 0.0 }
-
-// RateMean is a no-op.
-func (NilMeter) RateMean() float64 { return 0.0 }
-
-// Snapshot is a no-op.
-func (NilMeter) Snapshot() Meter { return NilMeter{} }
-
-// Stop is a no-op.
-func (NilMeter) Stop() {}
+func (NilMeter) Count() int64            { return 0 }
+func (NilMeter) Mark(n int64)            {}
+func (NilMeter) Snapshot() MeterSnapshot { return (*emptySnapshot)(nil) }
+func (NilMeter) Stop()                   {}
 
 // StandardMeter is the standard implementation of a Meter.
 type StandardMeter struct {
-	lock        sync.RWMutex
-	snapshot    *MeterSnapshot
+	count     atomic.Int64
+	uncounted atomic.Int64 // not yet added to the EWMAs
+	rateMean  atomic.Uint64
+
 	a1, a5, a15 EWMA
 	startTime   time.Time
 	stopped     atomic.Bool
@@ -184,7 +115,6 @@ type StandardMeter struct {
 
 func newStandardMeter() *StandardMeter {
 	return &StandardMeter{
-		snapshot:  &MeterSnapshot{},
 		a1:        NewEWMA1(),
 		a5:        NewEWMA5(),
 		a15:       NewEWMA15(),
@@ -194,97 +124,42 @@ func newStandardMeter() *StandardMeter {
 
 // Stop stops the meter, Mark() will be a no-op if you use it after being stopped.
 func (m *StandardMeter) Stop() {
-	stopped := m.stopped.Swap(true)
-	if !stopped {
+	if stopped := m.stopped.Swap(true); !stopped {
 		arbiter.Lock()
 		delete(arbiter.meters, m)
 		arbiter.Unlock()
 	}
 }
 
-// Count returns the number of events recorded.
-// It updates the meter to be as accurate as possible
-func (m *StandardMeter) Count() int64 {
-	m.lock.Lock()
-	defer m.lock.Unlock()
-	m.updateMeter()
-	return m.snapshot.count
-}
-
 // Mark records the occurrence of n events.
 func (m *StandardMeter) Mark(n int64) {
-	m.snapshot.temp.Add(n)
-}
-
-// Rate1 returns the one-minute moving average rate of events per second.
-func (m *StandardMeter) Rate1() float64 {
-	m.lock.RLock()
-	defer m.lock.RUnlock()
-	return m.snapshot.rate1
-}
-
-// Rate5 returns the five-minute moving average rate of events per second.
-func (m *StandardMeter) Rate5() float64 {
-	m.lock.RLock()
-	defer m.lock.RUnlock()
-	return m.snapshot.rate5
-}
-
-// Rate15 returns the fifteen-minute moving average rate of events per second.
-func (m *StandardMeter) Rate15() float64 {
-	m.lock.RLock()
-	defer m.lock.RUnlock()
-	return m.snapshot.rate15
-}
-
-// RateMean returns the meter's mean rate of events per second.
-func (m *StandardMeter) RateMean() float64 {
-	m.lock.RLock()
-	defer m.lock.RUnlock()
-	return m.snapshot.rateMean
+	m.uncounted.Add(n)
 }
 
 // Snapshot returns a read-only copy of the meter.
-func (m *StandardMeter) Snapshot() Meter {
-	m.lock.RLock()
-	snapshot := MeterSnapshot{
-		count:    m.snapshot.count,
-		rate1:    m.snapshot.rate1,
-		rate5:    m.snapshot.rate5,
-		rate15:   m.snapshot.rate15,
-		rateMean: m.snapshot.rateMean,
+func (m *StandardMeter) Snapshot() MeterSnapshot {
+	return &meterSnapshot{
+		count:    m.count.Load() + m.uncounted.Load(),
+		rate1:    m.a1.Snapshot().Rate(),
+		rate5:    m.a5.Snapshot().Rate(),
+		rate15:   m.a15.Snapshot().Rate(),
+		rateMean: math.Float64frombits(m.rateMean.Load()),
 	}
-	snapshot.temp.Store(m.snapshot.temp.Load())
-	m.lock.RUnlock()
-	return &snapshot
-}
-
-func (m *StandardMeter) updateSnapshot() {
-	// should run with write lock held on m.lock
-	snapshot := m.snapshot
-	snapshot.rate1 = m.a1.Rate()
-	snapshot.rate5 = m.a5.Rate()
-	snapshot.rate15 = m.a15.Rate()
-	snapshot.rateMean = float64(snapshot.count) / time.Since(m.startTime).Seconds()
 }
 
-func (m *StandardMeter) updateMeter() {
-	// should only run with write lock held on m.lock
-	n := m.snapshot.temp.Swap(0)
-	m.snapshot.count += n
+func (m *StandardMeter) tick() {
+	// Take the uncounted values, add to count
+	n := m.uncounted.Swap(0)
+	count := m.count.Add(n)
+	m.rateMean.Store(math.Float64bits(float64(count) / time.Since(m.startTime).Seconds()))
+	// Update the EWMA's internal state
 	m.a1.Update(n)
 	m.a5.Update(n)
 	m.a15.Update(n)
-}
-
-func (m *StandardMeter) tick() {
-	m.lock.Lock()
-	defer m.lock.Unlock()
-	m.updateMeter()
+	// And trigger them to calculate the rates
 	m.a1.Tick()
 	m.a5.Tick()
 	m.a15.Tick()
-	m.updateSnapshot()
 }
 
 // meterArbiter ticks meters every 5s from a single goroutine.
diff --git a/metrics/meter_test.go b/metrics/meter_test.go
index b3f6cb8c0c..019c4d765b 100644
--- a/metrics/meter_test.go
+++ b/metrics/meter_test.go
@@ -12,11 +12,17 @@ func BenchmarkMeter(b *testing.B) {
 		m.Mark(1)
 	}
 }
-
+func TestMeter(t *testing.T) {
+	m := NewMeter()
+	m.Mark(47)
+	if v := m.Snapshot().Count(); v != 47 {
+		t.Fatalf("have %d want %d", v, 47)
+	}
+}
 func TestGetOrRegisterMeter(t *testing.T) {
 	r := NewRegistry()
 	NewRegisteredMeter("foo", r).Mark(47)
-	if m := GetOrRegisterMeter("foo", r); m.Count() != 47 {
+	if m := GetOrRegisterMeter("foo", r).Snapshot(); m.Count() != 47 {
 		t.Fatal(m.Count())
 	}
 }
@@ -31,10 +37,10 @@ func TestMeterDecay(t *testing.T) {
 	ma.meters[m] = struct{}{}
 	m.Mark(1)
 	ma.tickMeters()
-	rateMean := m.RateMean()
+	rateMean := m.Snapshot().RateMean()
 	time.Sleep(100 * time.Millisecond)
 	ma.tickMeters()
-	if m.RateMean() >= rateMean {
+	if m.Snapshot().RateMean() >= rateMean {
 		t.Error("m.RateMean() didn't decrease")
 	}
 }
@@ -42,7 +48,7 @@ func TestMeterDecay(t *testing.T) {
 func TestMeterNonzero(t *testing.T) {
 	m := NewMeter()
 	m.Mark(3)
-	if count := m.Count(); count != 3 {
+	if count := m.Snapshot().Count(); count != 3 {
 		t.Errorf("m.Count(): 3 != %v\n", count)
 	}
 }
@@ -59,16 +65,8 @@ func TestMeterStop(t *testing.T) {
 	}
 }
 
-func TestMeterSnapshot(t *testing.T) {
-	m := NewMeter()
-	m.Mark(1)
-	if snapshot := m.Snapshot(); m.RateMean() != snapshot.RateMean() {
-		t.Fatal(snapshot)
-	}
-}
-
 func TestMeterZero(t *testing.T) {
-	m := NewMeter()
+	m := NewMeter().Snapshot()
 	if count := m.Count(); count != 0 {
 		t.Errorf("m.Count(): 0 != %v\n", count)
 	}
@@ -79,13 +77,13 @@ func TestMeterRepeat(t *testing.T) {
 	for i := 0; i < 101; i++ {
 		m.Mark(int64(i))
 	}
-	if count := m.Count(); count != 5050 {
+	if count := m.Snapshot().Count(); count != 5050 {
 		t.Errorf("m.Count(): 5050 != %v\n", count)
 	}
 	for i := 0; i < 101; i++ {
 		m.Mark(int64(i))
 	}
-	if count := m.Count(); count != 10100 {
+	if count := m.Snapshot().Count(); count != 10100 {
 		t.Errorf("m.Count(): 10100 != %v\n", count)
 	}
 }
diff --git a/metrics/metrics.go b/metrics/metrics.go
index c206f16924..97f03fa31d 100644
--- a/metrics/metrics.go
+++ b/metrics/metrics.go
@@ -85,6 +85,12 @@ var runtimeSamples = []metrics.Sample{
 	{Name: "/sched/latencies:seconds"}, // histogram
 }
 
+func ReadRuntimeStats() *runtimeStats {
+	r := new(runtimeStats)
+	readRuntimeStats(r)
+	return r
+}
+
 func readRuntimeStats(v *runtimeStats) {
 	metrics.Read(runtimeSamples)
 	for _, s := range runtimeSamples {
diff --git a/metrics/metrics_test.go b/metrics/metrics_test.go
index 534c44139b..2861d5f2ca 100644
--- a/metrics/metrics_test.go
+++ b/metrics/metrics_test.go
@@ -98,8 +98,8 @@ func Example() {
 	t.Time(func() { time.Sleep(10 * time.Millisecond) })
 	t.Update(1)
 
-	fmt.Println(c.Count())
-	fmt.Println(t.Min())
+	fmt.Println(c.Snapshot().Count())
+	fmt.Println(t.Snapshot().Min())
 	// Output: 17
 	// 1
 }
diff --git a/metrics/opentsdb.go b/metrics/opentsdb.go
index 4d2e209238..e81690f943 100644
--- a/metrics/opentsdb.go
+++ b/metrics/opentsdb.go
@@ -65,15 +65,15 @@ func (c *OpenTSDBConfig) writeRegistry(w io.Writer, now int64, shortHostname str
 	c.Registry.Each(func(name string, i interface{}) {
 		switch metric := i.(type) {
 		case Counter:
-			fmt.Fprintf(w, "put %s.%s.count %d %d host=%s\n", c.Prefix, name, now, metric.Count(), shortHostname)
+			fmt.Fprintf(w, "put %s.%s.count %d %d host=%s\n", c.Prefix, name, now, metric.Snapshot().Count(), shortHostname)
 		case CounterFloat64:
-			fmt.Fprintf(w, "put %s.%s.count %d %f host=%s\n", c.Prefix, name, now, metric.Count(), shortHostname)
+			fmt.Fprintf(w, "put %s.%s.count %d %f host=%s\n", c.Prefix, name, now, metric.Snapshot().Count(), shortHostname)
 		case Gauge:
-			fmt.Fprintf(w, "put %s.%s.value %d %d host=%s\n", c.Prefix, name, now, metric.Value(), shortHostname)
+			fmt.Fprintf(w, "put %s.%s.value %d %d host=%s\n", c.Prefix, name, now, metric.Snapshot().Value(), shortHostname)
 		case GaugeFloat64:
-			fmt.Fprintf(w, "put %s.%s.value %d %f host=%s\n", c.Prefix, name, now, metric.Value(), shortHostname)
+			fmt.Fprintf(w, "put %s.%s.value %d %f host=%s\n", c.Prefix, name, now, metric.Snapshot().Value(), shortHostname)
 		case GaugeInfo:
-			fmt.Fprintf(w, "put %s.%s.value %d %s host=%s\n", c.Prefix, name, now, metric.Value().String(), shortHostname)
+			fmt.Fprintf(w, "put %s.%s.value %d %s host=%s\n", c.Prefix, name, now, metric.Snapshot().Value().String(), shortHostname)
 		case Histogram:
 			h := metric.Snapshot()
 			ps := h.Percentiles([]float64{0.5, 0.75, 0.95, 0.99, 0.999})
diff --git a/metrics/opentsdb_test.go b/metrics/opentsdb_test.go
index c02b98af06..4548309f9c 100644
--- a/metrics/opentsdb_test.go
+++ b/metrics/opentsdb_test.go
@@ -1,6 +1,7 @@
 package metrics
 
 import (
+	"fmt"
 	"net"
 	"os"
 	"strings"
@@ -47,5 +48,19 @@ func TestExampleOpenTSB(t *testing.T) {
 	}
 	if have, want := w.String(), string(wantB); have != want {
 		t.Errorf("\nhave:\n%v\nwant:\n%v\n", have, want)
+		t.Logf("have vs want:\n%v", findFirstDiffPos(have, want))
 	}
 }
+
+func findFirstDiffPos(a, b string) string {
+	yy := strings.Split(b, "\n")
+	for i, x := range strings.Split(a, "\n") {
+		if i >= len(yy) {
+			return fmt.Sprintf("have:%d: %s\nwant:%d: <EOF>", i, x, i)
+		}
+		if y := yy[i]; x != y {
+			return fmt.Sprintf("have:%d: %s\nwant:%d: %s", i, x, i, y)
+		}
+	}
+	return ""
+}
diff --git a/metrics/prometheus/collector.go b/metrics/prometheus/collector.go
index 8624311c4b..25b258d56a 100644
--- a/metrics/prometheus/collector.go
+++ b/metrics/prometheus/collector.go
@@ -75,27 +75,27 @@ func (c *collector) Add(name string, i any) error {
 	return nil
 }
 
-func (c *collector) addCounter(name string, m metrics.Counter) {
+func (c *collector) addCounter(name string, m metrics.CounterSnapshot) {
 	c.writeGaugeCounter(name, m.Count())
 }
 
-func (c *collector) addCounterFloat64(name string, m metrics.CounterFloat64) {
+func (c *collector) addCounterFloat64(name string, m metrics.CounterFloat64Snapshot) {
 	c.writeGaugeCounter(name, m.Count())
 }
 
-func (c *collector) addGauge(name string, m metrics.Gauge) {
+func (c *collector) addGauge(name string, m metrics.GaugeSnapshot) {
 	c.writeGaugeCounter(name, m.Value())
 }
 
-func (c *collector) addGaugeFloat64(name string, m metrics.GaugeFloat64) {
+func (c *collector) addGaugeFloat64(name string, m metrics.GaugeFloat64Snapshot) {
 	c.writeGaugeCounter(name, m.Value())
 }
 
-func (c *collector) addGaugeInfo(name string, m metrics.GaugeInfo) {
+func (c *collector) addGaugeInfo(name string, m metrics.GaugeInfoSnapshot) {
 	c.writeGaugeInfo(name, m.Value())
 }
 
-func (c *collector) addHistogram(name string, m metrics.Histogram) {
+func (c *collector) addHistogram(name string, m metrics.HistogramSnapshot) {
 	pv := []float64{0.5, 0.75, 0.95, 0.99, 0.999, 0.9999}
 	ps := m.Percentiles(pv)
 	c.writeSummaryCounter(name, m.Count())
@@ -106,11 +106,11 @@ func (c *collector) addHistogram(name string, m metrics.Histogram) {
 	c.buff.WriteRune('\n')
 }
 
-func (c *collector) addMeter(name string, m metrics.Meter) {
+func (c *collector) addMeter(name string, m metrics.MeterSnapshot) {
 	c.writeGaugeCounter(name, m.Count())
 }
 
-func (c *collector) addTimer(name string, m metrics.Timer) {
+func (c *collector) addTimer(name string, m metrics.TimerSnapshot) {
 	pv := []float64{0.5, 0.75, 0.95, 0.99, 0.999, 0.9999}
 	ps := m.Percentiles(pv)
 	c.writeSummaryCounter(name, m.Count())
@@ -121,13 +121,12 @@ func (c *collector) addTimer(name string, m metrics.Timer) {
 	c.buff.WriteRune('\n')
 }
 
-func (c *collector) addResettingTimer(name string, m metrics.ResettingTimer) {
-	if len(m.Values()) <= 0 {
+func (c *collector) addResettingTimer(name string, m metrics.ResettingTimerSnapshot) {
+	if m.Count() <= 0 {
 		return
 	}
-	ps := m.Percentiles([]float64{50, 95, 99})
-	val := m.Values()
-	c.writeSummaryCounter(name, len(val))
+	ps := m.Percentiles([]float64{0.50, 0.95, 0.99})
+	c.writeSummaryCounter(name, m.Count())
 	c.buff.WriteString(fmt.Sprintf(typeSummaryTpl, mutateKey(name)))
 	c.writeSummaryPercentile(name, "0.50", ps[0])
 	c.writeSummaryPercentile(name, "0.95", ps[1])
diff --git a/metrics/prometheus/collector_test.go b/metrics/prometheus/collector_test.go
index 3d7903d4ad..ea17aac458 100644
--- a/metrics/prometheus/collector_test.go
+++ b/metrics/prometheus/collector_test.go
@@ -55,10 +55,10 @@ func findFirstDiffPos(a, b string) string {
 	yy := strings.Split(b, "\n")
 	for i, x := range strings.Split(a, "\n") {
 		if i >= len(yy) {
-			return fmt.Sprintf("a:%d: %s\nb:%d: <EOF>", i, x, i)
+			return fmt.Sprintf("have:%d: %s\nwant:%d: <EOF>", i, x, i)
 		}
 		if y := yy[i]; x != y {
-			return fmt.Sprintf("a:%d: %s\nb:%d: %s", i, x, i, y)
+			return fmt.Sprintf("have:%d: %s\nwant:%d: %s", i, x, i, y)
 		}
 	}
 	return ""
diff --git a/metrics/prometheus/testdata/prometheus.want b/metrics/prometheus/testdata/prometheus.want
index f35496e61d..861c5f5cf0 100644
--- a/metrics/prometheus/testdata/prometheus.want
+++ b/metrics/prometheus/testdata/prometheus.want
@@ -1,3 +1,25 @@
+# TYPE system_cpu_schedlatency_count counter
+system_cpu_schedlatency_count 5645
+
+# TYPE system_cpu_schedlatency summary
+system_cpu_schedlatency {quantile="0.5"} 0
+system_cpu_schedlatency {quantile="0.75"} 7168
+system_cpu_schedlatency {quantile="0.95"} 1.6777216e+07
+system_cpu_schedlatency {quantile="0.99"} 2.9360128e+07
+system_cpu_schedlatency {quantile="0.999"} 3.3554432e+07
+system_cpu_schedlatency {quantile="0.9999"} 3.3554432e+07
+
+# TYPE system_memory_pauses_count counter
+system_memory_pauses_count 14
+
+# TYPE system_memory_pauses summary
+system_memory_pauses {quantile="0.5"} 32768
+system_memory_pauses {quantile="0.75"} 57344
+system_memory_pauses {quantile="0.95"} 196608
+system_memory_pauses {quantile="0.99"} 196608
+system_memory_pauses {quantile="0.999"} 196608
+system_memory_pauses {quantile="0.9999"} 196608
+
 # TYPE test_counter gauge
 test_counter 12345
 
@@ -31,9 +53,9 @@ test_meter 0
 test_resetting_timer_count 6
 
 # TYPE test_resetting_timer summary
-test_resetting_timer {quantile="0.50"} 12000000
-test_resetting_timer {quantile="0.95"} 120000000
-test_resetting_timer {quantile="0.99"} 120000000
+test_resetting_timer {quantile="0.50"} 1.25e+07
+test_resetting_timer {quantile="0.95"} 1.2e+08
+test_resetting_timer {quantile="0.99"} 1.2e+08
 
 # TYPE test_timer_count counter
 test_timer_count 6
diff --git a/metrics/registry.go b/metrics/registry.go
index 66dbc890c0..8bfbc08042 100644
--- a/metrics/registry.go
+++ b/metrics/registry.go
@@ -150,13 +150,13 @@ func (r *StandardRegistry) GetAll() map[string]map[string]interface{} {
 		values := make(map[string]interface{})
 		switch metric := i.(type) {
 		case Counter:
-			values["count"] = metric.Count()
+			values["count"] = metric.Snapshot().Count()
 		case CounterFloat64:
-			values["count"] = metric.Count()
+			values["count"] = metric.Snapshot().Count()
 		case Gauge:
-			values["value"] = metric.Value()
+			values["value"] = metric.Snapshot().Value()
 		case GaugeFloat64:
-			values["value"] = metric.Value()
+			values["value"] = metric.Snapshot().Value()
 		case Healthcheck:
 			values["error"] = nil
 			metric.Check()
diff --git a/metrics/registry_test.go b/metrics/registry_test.go
index 7cc5cf14fe..75012dd4ac 100644
--- a/metrics/registry_test.go
+++ b/metrics/registry_test.go
@@ -85,11 +85,11 @@ func TestRegistryDuplicate(t *testing.T) {
 func TestRegistryGet(t *testing.T) {
 	r := NewRegistry()
 	r.Register("foo", NewCounter())
-	if count := r.Get("foo").(Counter).Count(); count != 0 {
+	if count := r.Get("foo").(Counter).Snapshot().Count(); count != 0 {
 		t.Fatal(count)
 	}
 	r.Get("foo").(Counter).Inc(1)
-	if count := r.Get("foo").(Counter).Count(); count != 1 {
+	if count := r.Get("foo").(Counter).Snapshot().Count(); count != 1 {
 		t.Fatal(count)
 	}
 }
diff --git a/metrics/resetting_sample.go b/metrics/resetting_sample.go
index 43c1129cd0..c38ffcd3ec 100644
--- a/metrics/resetting_sample.go
+++ b/metrics/resetting_sample.go
@@ -17,7 +17,7 @@ type resettingSample struct {
 }
 
 // Snapshot returns a read-only copy of the sample with the original reset.
-func (rs *resettingSample) Snapshot() Sample {
+func (rs *resettingSample) Snapshot() SampleSnapshot {
 	s := rs.Sample.Snapshot()
 	rs.Sample.Clear()
 	return s
diff --git a/metrics/resetting_timer.go b/metrics/resetting_timer.go
index 8e23c8eeea..6802e3fcea 100644
--- a/metrics/resetting_timer.go
+++ b/metrics/resetting_timer.go
@@ -1,22 +1,24 @@
 package metrics
 
 import (
-	"math"
 	"sync"
 	"time"
-
-	"golang.org/x/exp/slices"
 )
 
 // Initial slice capacity for the values stored in a ResettingTimer
 const InitialResettingTimerSliceCap = 10
 
+type ResettingTimerSnapshot interface {
+	Count() int
+	Mean() float64
+	Max() int64
+	Min() int64
+	Percentiles([]float64) []float64
+}
+
 // ResettingTimer is used for storing aggregated values for timers, which are reset on every flush interval.
 type ResettingTimer interface {
-	Values() []int64
-	Snapshot() ResettingTimer
-	Percentiles([]float64) []int64
-	Mean() float64
+	Snapshot() ResettingTimerSnapshot
 	Time(func())
 	Update(time.Duration)
 	UpdateSince(time.Time)
@@ -52,70 +54,40 @@ func NewResettingTimer() ResettingTimer {
 }
 
 // NilResettingTimer is a no-op ResettingTimer.
-type NilResettingTimer struct {
-}
-
-// Values is a no-op.
-func (NilResettingTimer) Values() []int64 { return nil }
-
-// Snapshot is a no-op.
-func (NilResettingTimer) Snapshot() ResettingTimer {
-	return &ResettingTimerSnapshot{
-		values: []int64{},
-	}
-}
-
-// Time is a no-op.
-func (NilResettingTimer) Time(f func()) { f() }
-
-// Update is a no-op.
-func (NilResettingTimer) Update(time.Duration) {}
-
-// Percentiles panics.
-func (NilResettingTimer) Percentiles([]float64) []int64 {
-	panic("Percentiles called on a NilResettingTimer")
-}
-
-// Mean panics.
-func (NilResettingTimer) Mean() float64 {
-	panic("Mean called on a NilResettingTimer")
-}
-
-// UpdateSince is a no-op.
-func (NilResettingTimer) UpdateSince(time.Time) {}
+type NilResettingTimer struct{}
+
+func (NilResettingTimer) Values() []int64                    { return nil }
+func (n NilResettingTimer) Snapshot() ResettingTimerSnapshot { return n }
+func (NilResettingTimer) Time(f func())                      { f() }
+func (NilResettingTimer) Update(time.Duration)               {}
+func (NilResettingTimer) Percentiles([]float64) []float64    { return nil }
+func (NilResettingTimer) Mean() float64                      { return 0.0 }
+func (NilResettingTimer) Max() int64                         { return 0 }
+func (NilResettingTimer) Min() int64                         { return 0 }
+func (NilResettingTimer) UpdateSince(time.Time)              {}
+func (NilResettingTimer) Count() int                         { return 0 }
 
 // StandardResettingTimer is the standard implementation of a ResettingTimer.
 // and Meter.
 type StandardResettingTimer struct {
 	values []int64
-	mutex  sync.Mutex
-}
+	sum    int64 // sum is a running count of the total sum, used later to calculate mean
 
-// Values returns a slice with all measurements.
-func (t *StandardResettingTimer) Values() []int64 {
-	return t.values
+	mutex sync.Mutex
 }
 
 // Snapshot resets the timer and returns a read-only copy of its contents.
-func (t *StandardResettingTimer) Snapshot() ResettingTimer {
+func (t *StandardResettingTimer) Snapshot() ResettingTimerSnapshot {
 	t.mutex.Lock()
 	defer t.mutex.Unlock()
-	currentValues := t.values
-	t.values = make([]int64, 0, InitialResettingTimerSliceCap)
-
-	return &ResettingTimerSnapshot{
-		values: currentValues,
+	snapshot := &resettingTimerSnapshot{}
+	if len(t.values) > 0 {
+		snapshot.mean = float64(t.sum) / float64(len(t.values))
+		snapshot.values = t.values
+		t.values = make([]int64, 0, InitialResettingTimerSliceCap)
 	}
-}
-
-// Percentiles panics.
-func (t *StandardResettingTimer) Percentiles([]float64) []int64 {
-	panic("Percentiles called on a StandardResettingTimer")
-}
-
-// Mean panics.
-func (t *StandardResettingTimer) Mean() float64 {
-	panic("Mean called on a StandardResettingTimer")
+	t.sum = 0
+	return snapshot
 }
 
 // Record the duration of the execution of the given function.
@@ -130,106 +102,70 @@ func (t *StandardResettingTimer) Update(d time.Duration) {
 	t.mutex.Lock()
 	defer t.mutex.Unlock()
 	t.values = append(t.values, int64(d))
+	t.sum += int64(d)
 }
 
 // Record the duration of an event that started at a time and ends now.
 func (t *StandardResettingTimer) UpdateSince(ts time.Time) {
-	t.mutex.Lock()
-	defer t.mutex.Unlock()
-	t.values = append(t.values, int64(time.Since(ts)))
+	t.Update(time.Since(ts))
 }
 
-// ResettingTimerSnapshot is a point-in-time copy of another ResettingTimer.
-type ResettingTimerSnapshot struct {
+// resettingTimerSnapshot is a point-in-time copy of another ResettingTimer.
+type resettingTimerSnapshot struct {
 	values              []int64
 	mean                float64
-	thresholdBoundaries []int64
+	max                 int64
+	min                 int64
+	thresholdBoundaries []float64
 	calculated          bool
 }
 
-// Snapshot returns the snapshot.
-func (t *ResettingTimerSnapshot) Snapshot() ResettingTimer { return t }
-
-// Time panics.
-func (*ResettingTimerSnapshot) Time(func()) {
-	panic("Time called on a ResettingTimerSnapshot")
-}
-
-// Update panics.
-func (*ResettingTimerSnapshot) Update(time.Duration) {
-	panic("Update called on a ResettingTimerSnapshot")
-}
-
-// UpdateSince panics.
-func (*ResettingTimerSnapshot) UpdateSince(time.Time) {
-	panic("UpdateSince called on a ResettingTimerSnapshot")
-}
-
-// Values returns all values from snapshot.
-func (t *ResettingTimerSnapshot) Values() []int64 {
-	return t.values
+// Count return the length of the values from snapshot.
+func (t *resettingTimerSnapshot) Count() int {
+	return len(t.values)
 }
 
 // Percentiles returns the boundaries for the input percentiles.
-func (t *ResettingTimerSnapshot) Percentiles(percentiles []float64) []int64 {
+// note: this method is not thread safe
+func (t *resettingTimerSnapshot) Percentiles(percentiles []float64) []float64 {
 	t.calc(percentiles)
-
 	return t.thresholdBoundaries
 }
 
 // Mean returns the mean of the snapshotted values
-func (t *ResettingTimerSnapshot) Mean() float64 {
+// note: this method is not thread safe
+func (t *resettingTimerSnapshot) Mean() float64 {
 	if !t.calculated {
-		t.calc([]float64{})
+		t.calc(nil)
 	}
 
 	return t.mean
 }
 
-func (t *ResettingTimerSnapshot) calc(percentiles []float64) {
-	slices.Sort(t.values)
-
-	count := len(t.values)
-	if count > 0 {
-		min := t.values[0]
-		max := t.values[count-1]
-
-		cumulativeValues := make([]int64, count)
-		cumulativeValues[0] = min
-		for i := 1; i < count; i++ {
-			cumulativeValues[i] = t.values[i] + cumulativeValues[i-1]
-		}
-
-		t.thresholdBoundaries = make([]int64, len(percentiles))
-
-		thresholdBoundary := max
-
-		for i, pct := range percentiles {
-			if count > 1 {
-				var abs float64
-				if pct >= 0 {
-					abs = pct
-				} else {
-					abs = 100 + pct
-				}
-				// poor man's math.Round(x):
-				// math.Floor(x + 0.5)
-				indexOfPerc := int(math.Floor(((abs / 100.0) * float64(count)) + 0.5))
-				if pct >= 0 && indexOfPerc > 0 {
-					indexOfPerc -= 1 // index offset=0
-				}
-				thresholdBoundary = t.values[indexOfPerc]
-			}
-
-			t.thresholdBoundaries[i] = thresholdBoundary
-		}
-
-		sum := cumulativeValues[count-1]
-		t.mean = float64(sum) / float64(count)
-	} else {
-		t.thresholdBoundaries = make([]int64, len(percentiles))
-		t.mean = 0
+// Max returns the max of the snapshotted values
+// note: this method is not thread safe
+func (t *resettingTimerSnapshot) Max() int64 {
+	if !t.calculated {
+		t.calc(nil)
+	}
+	return t.max
+}
+
+// Min returns the min of the snapshotted values
+// note: this method is not thread safe
+func (t *resettingTimerSnapshot) Min() int64 {
+	if !t.calculated {
+		t.calc(nil)
 	}
+	return t.min
+}
 
-	t.calculated = true
+func (t *resettingTimerSnapshot) calc(percentiles []float64) {
+	scores := CalculatePercentiles(t.values, percentiles)
+	t.thresholdBoundaries = scores
+	if len(t.values) == 0 {
+		return
+	}
+	t.min = t.values[0]
+	t.max = t.values[len(t.values)-1]
 }
diff --git a/metrics/resetting_timer_test.go b/metrics/resetting_timer_test.go
index 77c49dc386..4571fc8eb0 100644
--- a/metrics/resetting_timer_test.go
+++ b/metrics/resetting_timer_test.go
@@ -10,9 +10,9 @@ func TestResettingTimer(t *testing.T) {
 		values   []int64
 		start    int
 		end      int
-		wantP50  int64
-		wantP95  int64
-		wantP99  int64
+		wantP50  float64
+		wantP95  float64
+		wantP99  float64
 		wantMean float64
 		wantMin  int64
 		wantMax  int64
@@ -21,14 +21,14 @@ func TestResettingTimer(t *testing.T) {
 			values:  []int64{},
 			start:   1,
 			end:     11,
-			wantP50: 5, wantP95: 10, wantP99: 10,
+			wantP50: 5.5, wantP95: 10, wantP99: 10,
 			wantMin: 1, wantMax: 10, wantMean: 5.5,
 		},
 		{
 			values:  []int64{},
 			start:   1,
 			end:     101,
-			wantP50: 50, wantP95: 95, wantP99: 99,
+			wantP50: 50.5, wantP95: 95.94999999999999, wantP99: 99.99,
 			wantMin: 1, wantMax: 100, wantMean: 50.5,
 		},
 		{
@@ -56,11 +56,11 @@ func TestResettingTimer(t *testing.T) {
 			values:  []int64{1, 10},
 			start:   0,
 			end:     0,
-			wantP50: 1, wantP95: 10, wantP99: 10,
+			wantP50: 5.5, wantP95: 10, wantP99: 10,
 			wantMin: 1, wantMax: 10, wantMean: 5.5,
 		},
 	}
-	for ind, tt := range tests {
+	for i, tt := range tests {
 		timer := NewResettingTimer()
 
 		for i := tt.start; i < tt.end; i++ {
@@ -70,37 +70,27 @@ func TestResettingTimer(t *testing.T) {
 		for _, v := range tt.values {
 			timer.Update(time.Duration(v))
 		}
-
 		snap := timer.Snapshot()
 
-		ps := snap.Percentiles([]float64{50, 95, 99})
-
-		val := snap.Values()
+		ps := snap.Percentiles([]float64{0.50, 0.95, 0.99})
 
-		if len(val) > 0 {
-			if tt.wantMin != val[0] {
-				t.Fatalf("%d: min: got %d, want %d", ind, val[0], tt.wantMin)
-			}
-
-			if tt.wantMax != val[len(val)-1] {
-				t.Fatalf("%d: max: got %d, want %d", ind, val[len(val)-1], tt.wantMax)
-			}
+		if have, want := snap.Min(), tt.wantMin; have != want {
+			t.Fatalf("%d: min: have %d, want %d", i, have, want)
 		}
-
-		if tt.wantMean != snap.Mean() {
-			t.Fatalf("%d: mean: got %.2f, want %.2f", ind, snap.Mean(), tt.wantMean)
+		if have, want := snap.Max(), tt.wantMax; have != want {
+			t.Fatalf("%d: max: have %d, want %d", i, have, want)
 		}
-
-		if tt.wantP50 != ps[0] {
-			t.Fatalf("%d: p50: got %d, want %d", ind, ps[0], tt.wantP50)
+		if have, want := snap.Mean(), tt.wantMean; have != want {
+			t.Fatalf("%d: mean: have %v, want %v", i, have, want)
 		}
-
-		if tt.wantP95 != ps[1] {
-			t.Fatalf("%d: p95: got %d, want %d", ind, ps[1], tt.wantP95)
+		if have, want := ps[0], tt.wantP50; have != want {
+			t.Errorf("%d: p50: have %v, want %v", i, have, want)
 		}
-
-		if tt.wantP99 != ps[2] {
-			t.Fatalf("%d: p99: got %d, want %d", ind, ps[2], tt.wantP99)
+		if have, want := ps[1], tt.wantP95; have != want {
+			t.Errorf("%d: p95: have %v, want %v", i, have, want)
+		}
+		if have, want := ps[2], tt.wantP99; have != want {
+			t.Errorf("%d: p99: have %v, want %v", i, have, want)
 		}
 	}
 }
@@ -110,11 +100,11 @@ func TestResettingTimerWithFivePercentiles(t *testing.T) {
 		values   []int64
 		start    int
 		end      int
-		wantP05  int64
-		wantP20  int64
-		wantP50  int64
-		wantP95  int64
-		wantP99  int64
+		wantP05  float64
+		wantP20  float64
+		wantP50  float64
+		wantP95  float64
+		wantP99  float64
 		wantMean float64
 		wantMin  int64
 		wantMax  int64
@@ -123,14 +113,14 @@ func TestResettingTimerWithFivePercentiles(t *testing.T) {
 			values:  []int64{},
 			start:   1,
 			end:     11,
-			wantP05: 1, wantP20: 2, wantP50: 5, wantP95: 10, wantP99: 10,
+			wantP05: 1, wantP20: 2.2, wantP50: 5.5, wantP95: 10, wantP99: 10,
 			wantMin: 1, wantMax: 10, wantMean: 5.5,
 		},
 		{
 			values:  []int64{},
 			start:   1,
 			end:     101,
-			wantP05: 5, wantP20: 20, wantP50: 50, wantP95: 95, wantP99: 99,
+			wantP05: 5.050000000000001, wantP20: 20.200000000000003, wantP50: 50.5, wantP95: 95.94999999999999, wantP99: 99.99,
 			wantMin: 1, wantMax: 100, wantMean: 50.5,
 		},
 		{
@@ -158,7 +148,7 @@ func TestResettingTimerWithFivePercentiles(t *testing.T) {
 			values:  []int64{1, 10},
 			start:   0,
 			end:     0,
-			wantP05: 1, wantP20: 1, wantP50: 1, wantP95: 10, wantP99: 10,
+			wantP05: 1, wantP20: 1, wantP50: 5.5, wantP95: 10, wantP99: 10,
 			wantMin: 1, wantMax: 10, wantMean: 5.5,
 		},
 	}
@@ -175,42 +165,33 @@ func TestResettingTimerWithFivePercentiles(t *testing.T) {
 
 		snap := timer.Snapshot()
 
-		ps := snap.Percentiles([]float64{5, 20, 50, 95, 99})
-
-		val := snap.Values()
+		ps := snap.Percentiles([]float64{0.05, 0.20, 0.50, 0.95, 0.99})
 
-		if len(val) > 0 {
-			if tt.wantMin != val[0] {
-				t.Fatalf("%d: min: got %d, want %d", ind, val[0], tt.wantMin)
-			}
+		if tt.wantMin != snap.Min() {
+			t.Errorf("%d: min: got %d, want %d", ind, snap.Min(), tt.wantMin)
+		}
 
-			if tt.wantMax != val[len(val)-1] {
-				t.Fatalf("%d: max: got %d, want %d", ind, val[len(val)-1], tt.wantMax)
-			}
+		if tt.wantMax != snap.Max() {
+			t.Errorf("%d: max: got %d, want %d", ind, snap.Max(), tt.wantMax)
 		}
 
 		if tt.wantMean != snap.Mean() {
-			t.Fatalf("%d: mean: got %.2f, want %.2f", ind, snap.Mean(), tt.wantMean)
+			t.Errorf("%d: mean: got %.2f, want %.2f", ind, snap.Mean(), tt.wantMean)
 		}
-
 		if tt.wantP05 != ps[0] {
-			t.Fatalf("%d: p05: got %d, want %d", ind, ps[0], tt.wantP05)
+			t.Errorf("%d: p05: got %v, want %v", ind, ps[0], tt.wantP05)
 		}
-
 		if tt.wantP20 != ps[1] {
-			t.Fatalf("%d: p20: got %d, want %d", ind, ps[1], tt.wantP20)
+			t.Errorf("%d: p20: got %v, want %v", ind, ps[1], tt.wantP20)
 		}
-
 		if tt.wantP50 != ps[2] {
-			t.Fatalf("%d: p50: got %d, want %d", ind, ps[2], tt.wantP50)
+			t.Errorf("%d: p50: got %v, want %v", ind, ps[2], tt.wantP50)
 		}
-
 		if tt.wantP95 != ps[3] {
-			t.Fatalf("%d: p95: got %d, want %d", ind, ps[3], tt.wantP95)
+			t.Errorf("%d: p95: got %v, want %v", ind, ps[3], tt.wantP95)
 		}
-
 		if tt.wantP99 != ps[4] {
-			t.Fatalf("%d: p99: got %d, want %d", ind, ps[4], tt.wantP99)
+			t.Errorf("%d: p99: got %v, want %v", ind, ps[4], tt.wantP99)
 		}
 	}
 }
diff --git a/metrics/runtimehistogram.go b/metrics/runtimehistogram.go
index c68939af1e..92fcbcc281 100644
--- a/metrics/runtimehistogram.go
+++ b/metrics/runtimehistogram.go
@@ -17,13 +17,19 @@ func getOrRegisterRuntimeHistogram(name string, scale float64, r Registry) *runt
 
 // runtimeHistogram wraps a runtime/metrics histogram.
 type runtimeHistogram struct {
-	v           atomic.Value
+	v           atomic.Value // v is a pointer to a metrics.Float64Histogram
 	scaleFactor float64
 }
 
 func newRuntimeHistogram(scale float64) *runtimeHistogram {
 	h := &runtimeHistogram{scaleFactor: scale}
-	h.update(&metrics.Float64Histogram{})
+	h.update(new(metrics.Float64Histogram))
+	return h
+}
+
+func RuntimeHistogramFromData(scale float64, hist *metrics.Float64Histogram) *runtimeHistogram {
+	h := &runtimeHistogram{scaleFactor: scale}
+	h.update(hist)
 	return h
 }
 
@@ -35,130 +41,107 @@ func (h *runtimeHistogram) update(mh *metrics.Float64Histogram) {
 		return
 	}
 
-	s := runtimeHistogramSnapshot{
+	s := metrics.Float64Histogram{
 		Counts:  make([]uint64, len(mh.Counts)),
 		Buckets: make([]float64, len(mh.Buckets)),
 	}
 	copy(s.Counts, mh.Counts)
-	copy(s.Buckets, mh.Buckets)
-	for i, b := range s.Buckets {
+	for i, b := range mh.Buckets {
 		s.Buckets[i] = b * h.scaleFactor
 	}
 	h.v.Store(&s)
 }
 
-func (h *runtimeHistogram) load() *runtimeHistogramSnapshot {
-	return h.v.Load().(*runtimeHistogramSnapshot)
-}
-
 func (h *runtimeHistogram) Clear() {
 	panic("runtimeHistogram does not support Clear")
 }
 func (h *runtimeHistogram) Update(int64) {
 	panic("runtimeHistogram does not support Update")
 }
-func (h *runtimeHistogram) Sample() Sample {
-	return NilSample{}
-}
-
-// Snapshot returns a non-changing cop of the histogram.
-func (h *runtimeHistogram) Snapshot() Histogram {
-	return h.load()
-}
-
-// Count returns the sample count.
-func (h *runtimeHistogram) Count() int64 {
-	return h.load().Count()
-}
-
-// Mean returns an approximation of the mean.
-func (h *runtimeHistogram) Mean() float64 {
-	return h.load().Mean()
-}
-
-// StdDev approximates the standard deviation of the histogram.
-func (h *runtimeHistogram) StdDev() float64 {
-	return h.load().StdDev()
-}
-
-// Variance approximates the variance of the histogram.
-func (h *runtimeHistogram) Variance() float64 {
-	return h.load().Variance()
-}
-
-// Percentile computes the p'th percentile value.
-func (h *runtimeHistogram) Percentile(p float64) float64 {
-	return h.load().Percentile(p)
-}
 
-// Percentiles computes all requested percentile values.
-func (h *runtimeHistogram) Percentiles(ps []float64) []float64 {
-	return h.load().Percentiles(ps)
-}
-
-// Max returns the highest sample value.
-func (h *runtimeHistogram) Max() int64 {
-	return h.load().Max()
+// Snapshot returns a non-changing copy of the histogram.
+func (h *runtimeHistogram) Snapshot() HistogramSnapshot {
+	hist := h.v.Load().(*metrics.Float64Histogram)
+	return newRuntimeHistogramSnapshot(hist)
 }
 
-// Min returns the lowest sample value.
-func (h *runtimeHistogram) Min() int64 {
-	return h.load().Min()
-}
-
-// Sum returns the sum of all sample values.
-func (h *runtimeHistogram) Sum() int64 {
-	return h.load().Sum()
+type runtimeHistogramSnapshot struct {
+	internal   *metrics.Float64Histogram
+	calculated bool
+	// The following fields are (lazily) calculated based on 'internal'
+	mean     float64
+	count    int64
+	min      int64 // min is the lowest sample value.
+	max      int64 // max is the highest sample value.
+	variance float64
 }
 
-type runtimeHistogramSnapshot metrics.Float64Histogram
-
-func (h *runtimeHistogramSnapshot) Clear() {
-	panic("runtimeHistogram does not support Clear")
-}
-func (h *runtimeHistogramSnapshot) Update(int64) {
-	panic("runtimeHistogram does not support Update")
-}
-func (h *runtimeHistogramSnapshot) Sample() Sample {
-	return NilSample{}
+func newRuntimeHistogramSnapshot(h *metrics.Float64Histogram) *runtimeHistogramSnapshot {
+	return &runtimeHistogramSnapshot{
+		internal: h,
+	}
 }
 
-func (h *runtimeHistogramSnapshot) Snapshot() Histogram {
-	return h
+// calc calculates the values for the snapshot. This method is not threadsafe.
+func (h *runtimeHistogramSnapshot) calc() {
+	h.calculated = true
+	var (
+		count int64   // number of samples
+		sum   float64 // approx sum of all sample values
+		min   int64
+		max   float64
+	)
+	if len(h.internal.Counts) == 0 {
+		return
+	}
+	for i, c := range h.internal.Counts {
+		if c == 0 {
+			continue
+		}
+		if count == 0 { // Set min only first loop iteration
+			min = int64(math.Floor(h.internal.Buckets[i]))
+		}
+		count += int64(c)
+		sum += h.midpoint(i) * float64(c)
+		// Set max on every iteration
+		edge := h.internal.Buckets[i+1]
+		if math.IsInf(edge, 1) {
+			edge = h.internal.Buckets[i]
+		}
+		if edge > max {
+			max = edge
+		}
+	}
+	h.min = min
+	h.max = int64(max)
+	h.mean = sum / float64(count)
+	h.count = count
 }
 
 // Count returns the sample count.
 func (h *runtimeHistogramSnapshot) Count() int64 {
-	var count int64
-	for _, c := range h.Counts {
-		count += int64(c)
+	if !h.calculated {
+		h.calc()
 	}
-	return count
+	return h.count
 }
 
-// Mean returns an approximation of the mean.
-func (h *runtimeHistogramSnapshot) Mean() float64 {
-	if len(h.Counts) == 0 {
-		return 0
-	}
-	mean, _ := h.mean()
-	return mean
+// Size returns the size of the sample at the time the snapshot was taken.
+func (h *runtimeHistogramSnapshot) Size() int {
+	return len(h.internal.Counts)
 }
 
-// mean computes the mean and also the total sample count.
-func (h *runtimeHistogramSnapshot) mean() (mean, totalCount float64) {
-	var sum float64
-	for i, c := range h.Counts {
-		midpoint := h.midpoint(i)
-		sum += midpoint * float64(c)
-		totalCount += float64(c)
+// Mean returns an approximation of the mean.
+func (h *runtimeHistogramSnapshot) Mean() float64 {
+	if !h.calculated {
+		h.calc()
 	}
-	return sum / totalCount, totalCount
+	return h.mean
 }
 
 func (h *runtimeHistogramSnapshot) midpoint(bucket int) float64 {
-	high := h.Buckets[bucket+1]
-	low := h.Buckets[bucket]
+	high := h.internal.Buckets[bucket+1]
+	low := h.internal.Buckets[bucket]
 	if math.IsInf(high, 1) {
 		// The edge of the highest bucket can be +Inf, and it's supposed to mean that this
 		// bucket contains all remaining samples > low. We can't get the middle of an
@@ -180,23 +163,31 @@ func (h *runtimeHistogramSnapshot) StdDev() float64 {
 
 // Variance approximates the variance of the histogram.
 func (h *runtimeHistogramSnapshot) Variance() float64 {
-	if len(h.Counts) == 0 {
+	if len(h.internal.Counts) == 0 {
 		return 0
 	}
-
-	mean, totalCount := h.mean()
-	if totalCount <= 1 {
+	if !h.calculated {
+		h.calc()
+	}
+	if h.count <= 1 {
 		// There is no variance when there are zero or one items.
 		return 0
 	}
-
+	// Variance is not calculated in 'calc', because it requires a second iteration.
+	// Therefore we calculate it lazily in this method, triggered either by
+	// a direct call to Variance or via StdDev.
+	if h.variance != 0.0 {
+		return h.variance
+	}
 	var sum float64
-	for i, c := range h.Counts {
+
+	for i, c := range h.internal.Counts {
 		midpoint := h.midpoint(i)
-		d := midpoint - mean
+		d := midpoint - h.mean
 		sum += float64(c) * (d * d)
 	}
-	return sum / (totalCount - 1)
+	h.variance = sum / float64(h.count-1)
+	return h.variance
 }
 
 // Percentile computes the p'th percentile value.
@@ -231,11 +222,11 @@ func (h *runtimeHistogramSnapshot) Percentiles(ps []float64) []float64 {
 
 func (h *runtimeHistogramSnapshot) computePercentiles(thresh []float64) {
 	var totalCount float64
-	for i, count := range h.Counts {
+	for i, count := range h.internal.Counts {
 		totalCount += float64(count)
 
 		for len(thresh) > 0 && thresh[0] < totalCount {
-			thresh[0] = h.Buckets[i]
+			thresh[0] = h.internal.Buckets[i]
 			thresh = thresh[1:]
 		}
 		if len(thresh) == 0 {
@@ -250,34 +241,25 @@ func (h *runtimeHistogramSnapshot) computePercentiles(thresh []float64) {
 
 // Max returns the highest sample value.
 func (h *runtimeHistogramSnapshot) Max() int64 {
-	for i := len(h.Counts) - 1; i >= 0; i-- {
-		count := h.Counts[i]
-		if count > 0 {
-			edge := h.Buckets[i+1]
-			if math.IsInf(edge, 1) {
-				edge = h.Buckets[i]
-			}
-			return int64(math.Ceil(edge))
-		}
+	if !h.calculated {
+		h.calc()
 	}
-	return 0
+	return h.max
 }
 
 // Min returns the lowest sample value.
 func (h *runtimeHistogramSnapshot) Min() int64 {
-	for i, count := range h.Counts {
-		if count > 0 {
-			return int64(math.Floor(h.Buckets[i]))
-		}
+	if !h.calculated {
+		h.calc()
 	}
-	return 0
+	return h.min
 }
 
 // Sum returns the sum of all sample values.
 func (h *runtimeHistogramSnapshot) Sum() int64 {
 	var sum float64
-	for i := range h.Counts {
-		sum += h.Buckets[i] * float64(h.Counts[i])
+	for i := range h.internal.Counts {
+		sum += h.internal.Buckets[i] * float64(h.internal.Counts[i])
 	}
 	return int64(math.Ceil(sum))
 }
diff --git a/metrics/runtimehistogram_test.go b/metrics/runtimehistogram_test.go
index d53a014383..cf7e36420a 100644
--- a/metrics/runtimehistogram_test.go
+++ b/metrics/runtimehistogram_test.go
@@ -1,11 +1,14 @@
 package metrics
 
 import (
+	"bytes"
+	"encoding/gob"
 	"fmt"
 	"math"
 	"reflect"
 	"runtime/metrics"
 	"testing"
+	"time"
 )
 
 var _ Histogram = (*runtimeHistogram)(nil)
@@ -74,7 +77,7 @@ func TestRuntimeHistogramStats(t *testing.T) {
 
 	for i, test := range tests {
 		t.Run(fmt.Sprint(i), func(t *testing.T) {
-			s := runtimeHistogramSnapshot(test.h)
+			s := RuntimeHistogramFromData(1.0, &test.h).Snapshot()
 
 			if v := s.Count(); v != test.Count {
 				t.Errorf("Count() = %v, want %v", v, test.Count)
@@ -121,13 +124,39 @@ func approxEqual(x, y, ε float64) bool {
 // This test verifies that requesting Percentiles in unsorted order
 // returns them in the requested order.
 func TestRuntimeHistogramStatsPercentileOrder(t *testing.T) {
-	p := runtimeHistogramSnapshot{
+	s := RuntimeHistogramFromData(1.0, &metrics.Float64Histogram{
 		Counts:  []uint64{1, 1, 1, 1, 1, 1, 1, 1, 1, 1},
 		Buckets: []float64{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10},
-	}
-	result := p.Percentiles([]float64{1, 0.2, 0.5, 0.1, 0.2})
+	}).Snapshot()
+	result := s.Percentiles([]float64{1, 0.2, 0.5, 0.1, 0.2})
 	expected := []float64{10, 2, 5, 1, 2}
 	if !reflect.DeepEqual(result, expected) {
 		t.Fatal("wrong result:", result)
 	}
 }
+
+func BenchmarkRuntimeHistogramSnapshotRead(b *testing.B) {
+	var sLatency = "7\xff\x81\x03\x01\x01\x10Float64Histogram\x01\xff\x82\x00\x01\x02\x01\x06Counts\x01\xff\x84\x00\x01\aBuckets\x01\xff\x86\x00\x00\x00\x16\xff\x83\x02\x01\x01\b[]uint64\x01\xff\x84\x00\x01\x06\x00\x00\x17\xff\x85\x02\x01\x01\t[]float64\x01\xff\x86\x00\x01\b\x00\x00\xfe\x06T\xff\x82\x01\xff\xa2\x00\xfe\r\xef\x00\x01\x02\x02\x04\x05\x04\b\x15\x17 B?6.L;$!2) \x1a? \x190aH7FY6#\x190\x1d\x14\x10\x1b\r\t\x04\x03\x01\x01\x00\x03\x02\x00\x03\x05\x05\x02\x02\x06\x04\v\x06\n\x15\x18\x13'&.\x12=H/L&\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\xff\xa3\xfe\xf0\xff\x00\xf8\x95\xd6&\xe8\v.q>\xf8\x95\xd6&\xe8\v.\x81>\xf8\xdfA:\xdc\x11ʼn>\xf8\x95\xd6&\xe8\v.\x91>\xf8:\x8c0\xe2\x8ey\x95>\xf8\xdfA:\xdc\x11ř>\xf8\x84\xf7C֔\x10\x9e>\xf8\x95\xd6&\xe8\v.\xa1>\xf8:\x8c0\xe2\x8ey\xa5>\xf8\xdfA:\xdc\x11ũ>\xf8\x84\xf7C֔\x10\xae>\xf8\x95\xd6&\xe8\v.\xb1>\xf8:\x8c0\xe2\x8ey\xb5>\xf8\xdfA:\xdc\x11Ź>\xf8\x84\xf7C֔\x10\xbe>\xf8\x95\xd6&\xe8\v.\xc1>\xf8:\x8c0\xe2\x8ey\xc5>\xf8\xdfA:\xdc\x11\xc5\xc9>\xf8\x84\xf7C֔\x10\xce>\xf8\x95\xd6&\xe8\v.\xd1>\xf8:\x8c0\xe2\x8ey\xd5>\xf8\xdfA:\xdc\x11\xc5\xd9>\xf8\x84\xf7C֔\x10\xde>\xf8\x95\xd6&\xe8\v.\xe1>\xf8:\x8c0\xe2\x8ey\xe5>\xf8\xdfA:\xdc\x11\xc5\xe9>\xf8\x84\xf7C֔\x10\xee>\xf8\x95\xd6&\xe8\v.\xf1>\xf8:\x8c0\xe2\x8ey\xf5>\xf8\xdfA:\xdc\x11\xc5\xf9>\xf8\x84\xf7C֔\x10\xfe>\xf8\x95\xd6&\xe8\v.\x01?\xf8:\x8c0\xe2\x8ey\x05?\xf8\xdfA:\xdc\x11\xc5\t?\xf8\x84\xf7C֔\x10\x0e?\xf8\x95\xd6&\xe8\v.\x11?\xf8:\x8c0\xe2\x8ey\x15?\xf8\xdfA:\xdc\x11\xc5\x19?\xf8\x84\xf7C֔\x10\x1e?\xf8\x95\xd6&\xe8\v.!?\xf8:\x8c0\xe2\x8ey%?\xf8\xdfA:\xdc\x11\xc5)?\xf8\x84\xf7C֔\x10.?\xf8\x95\xd6&\xe8\v.1?\xf8:\x8c0\xe2\x8ey5?\xf8\xdfA:\xdc\x11\xc59?\xf8\x84\xf7C֔\x10>?\xf8\x95\xd6&\xe8\v.A?\xf8:\x8c0\xe2\x8eyE?\xf8\xdfA:\xdc\x11\xc5I?\xf8\x84\xf7C֔\x10N?\xf8\x95\xd6&\xe8\v.Q?\xf8:\x8c0\xe2\x8eyU?\xf8\xdfA:\xdc\x11\xc5Y?\xf8\x84\xf7C֔\x10^?\xf8\x95\xd6&\xe8\v.a?\xf8:\x8c0\xe2\x8eye?\xf8\xdfA:\xdc\x11\xc5i?\xf8\x84\xf7C֔\x10n?\xf8\x95\xd6&\xe8\v.q?\xf8:\x8c0\xe2\x8eyu?\xf8\xdfA:\xdc\x11\xc5y?\xf8\x84\xf7C֔\x10~?\xf8\x95\xd6&\xe8\v.\x81?\xf8:\x8c0\xe2\x8ey\x85?\xf8\xdfA:\xdc\x11ʼn?\xf8\x84\xf7C֔\x10\x8e?\xf8\x95\xd6&\xe8\v.\x91?\xf8:\x8c0\xe2\x8ey\x95?\xf8\xdfA:\xdc\x11ř?\xf8\x84\xf7C֔\x10\x9e?\xf8\x95\xd6&\xe8\v.\xa1?\xf8:\x8c0\xe2\x8ey\xa5?\xf8\xdfA:\xdc\x11ũ?\xf8\x84\xf7C֔\x10\xae?\xf8\x95\xd6&\xe8\v.\xb1?\xf8:\x8c0\xe2\x8ey\xb5?\xf8\xdfA:\xdc\x11Ź?\xf8\x84\xf7C֔\x10\xbe?\xf8\x95\xd6&\xe8\v.\xc1?\xf8:\x8c0\xe2\x8ey\xc5?\xf8\xdfA:\xdc\x11\xc5\xc9?\xf8\x84\xf7C֔\x10\xce?\xf8\x95\xd6&\xe8\v.\xd1?\xf8:\x8c0\xe2\x8ey\xd5?\xf8\xdfA:\xdc\x11\xc5\xd9?\xf8\x84\xf7C֔\x10\xde?\xf8\x95\xd6&\xe8\v.\xe1?\xf8:\x8c0\xe2\x8ey\xe5?\xf8\xdfA:\xdc\x11\xc5\xe9?\xf8\x84\xf7C֔\x10\xee?\xf8\x95\xd6&\xe8\v.\xf1?\xf8:\x8c0\xe2\x8ey\xf5?\xf8\xdfA:\xdc\x11\xc5\xf9?\xf8\x84\xf7C֔\x10\xfe?\xf8\x95\xd6&\xe8\v.\x01@\xf8:\x8c0\xe2\x8ey\x05@\xf8\xdfA:\xdc\x11\xc5\t@\xf8\x84\xf7C֔\x10\x0e@\xf8\x95\xd6&\xe8\v.\x11@\xf8:\x8c0\xe2\x8ey\x15@\xf8\xdfA:\xdc\x11\xc5\x19@\xf8\x84\xf7C֔\x10\x1e@\xf8\x95\xd6&\xe8\v.!@\xf8:\x8c0\xe2\x8ey%@\xf8\xdfA:\xdc\x11\xc5)@\xf8\x84\xf7C֔\x10.@\xf8\x95\xd6&\xe8\v.1@\xf8:\x8c0\xe2\x8ey5@\xf8\xdfA:\xdc\x11\xc59@\xf8\x84\xf7C֔\x10>@\xf8\x95\xd6&\xe8\v.A@\xf8:\x8c0\xe2\x8eyE@\xf8\xdfA:\xdc\x11\xc5I@\xf8\x84\xf7C֔\x10N@\xf8\x95\xd6&\xe8\v.Q@\xf8:\x8c0\xe2\x8eyU@\xf8\xdfA:\xdc\x11\xc5Y@\xf8\x84\xf7C֔\x10^@\xf8\x95\xd6&\xe8\v.a@\xf8:\x8c0\xe2\x8eye@\xf8\xdfA:\xdc\x11\xc5i@\xf8\x84\xf7C֔\x10n@\xf8\x95\xd6&\xe8\v.q@\xf8:\x8c0\xe2\x8eyu@\xf8\xdfA:\xdc\x11\xc5y@\xf8\x84\xf7C֔\x10~@\xf8\x95\xd6&\xe8\v.\x81@\xf8:\x8c0\xe2\x8ey\x85@\xf8\xdfA:\xdc\x11ʼn@\xf8\x84\xf7C֔\x10\x8e@\xf8\x95\xd6&\xe8\v.\x91@\xf8:\x8c0\xe2\x8ey\x95@\xf8\xdfA:\xdc\x11ř@\xf8\x84\xf7C֔\x10\x9e@\xf8\x95\xd6&\xe8\v.\xa1@\xf8:\x8c0\xe2\x8ey\xa5@\xf8\xdfA:\xdc\x11ũ@\xf8\x84\xf7C֔\x10\xae@\xf8\x95\xd6&\xe8\v.\xb1@\xf8:\x8c0\xe2\x8ey\xb5@\xf8\xdfA:\xdc\x11Ź@\xf8\x84\xf7C֔\x10\xbe@\xf8\x95\xd6&\xe8\v.\xc1@\xf8:\x8c0\xe2\x8ey\xc5@\xf8\xdfA:\xdc\x11\xc5\xc9@\xf8\x84\xf7C֔\x10\xce@\xf8\x95\xd6&\xe8\v.\xd1@\xf8:\x8c0\xe2\x8ey\xd5@\xf8\xdfA:\xdc\x11\xc5\xd9@\xf8\x84\xf7C֔\x10\xde@\xf8\x95\xd6&\xe8\v.\xe1@\xf8:\x8c0\xe2\x8ey\xe5@\xf8\xdfA:\xdc\x11\xc5\xe9@\xf8\x84\xf7C֔\x10\xee@\xf8\x95\xd6&\xe8\v.\xf1@\xf8:\x8c0\xe2\x8ey\xf5@\xf8\xdfA:\xdc\x11\xc5\xf9@\xf8\x84\xf7C֔\x10\xfe@\xf8\x95\xd6&\xe8\v.\x01A\xfe\xf0\x7f\x00"
+
+	dserialize := func(data string) *metrics.Float64Histogram {
+		var res metrics.Float64Histogram
+		if err := gob.NewDecoder(bytes.NewReader([]byte(data))).Decode(&res); err != nil {
+			panic(err)
+		}
+		return &res
+	}
+	latency := RuntimeHistogramFromData(float64(time.Second), dserialize(sLatency))
+	b.ResetTimer()
+	b.ReportAllocs()
+	for i := 0; i < b.N; i++ {
+		snap := latency.Snapshot()
+		// These are the fields that influxdb accesses
+		_ = snap.Count()
+		_ = snap.Max()
+		_ = snap.Mean()
+		_ = snap.Min()
+		_ = snap.StdDev()
+		_ = snap.Variance()
+		_ = snap.Percentiles([]float64{0.25, 0.5, 0.75, 0.95, 0.99, 0.999, 0.9999})
+	}
+}
diff --git a/metrics/sample.go b/metrics/sample.go
index 252a878f58..5398dd42d5 100644
--- a/metrics/sample.go
+++ b/metrics/sample.go
@@ -11,10 +11,7 @@ import (
 
 const rescaleThreshold = time.Hour
 
-// Samples maintain a statistically-significant selection of values from
-// a stream.
-type Sample interface {
-	Clear()
+type SampleSnapshot interface {
 	Count() int64
 	Max() int64
 	Mean() float64
@@ -22,14 +19,19 @@ type Sample interface {
 	Percentile(float64) float64
 	Percentiles([]float64) []float64
 	Size() int
-	Snapshot() Sample
 	StdDev() float64
 	Sum() int64
-	Update(int64)
-	Values() []int64
 	Variance() float64
 }
 
+// Samples maintain a statistically-significant selection of values from
+// a stream.
+type Sample interface {
+	Snapshot() SampleSnapshot
+	Clear()
+	Update(int64)
+}
+
 // ExpDecaySample is an exponentially-decaying sample using a forward-decaying
 // priority reservoir.  See Cormode et al's "Forward Decay: A Practical Time
 // Decay Model for Streaming Systems".
@@ -77,72 +79,29 @@ func (s *ExpDecaySample) Clear() {
 	s.values.Clear()
 }
 
-// Count returns the number of samples recorded, which may exceed the
-// reservoir size.
-func (s *ExpDecaySample) Count() int64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return s.count
-}
-
-// Max returns the maximum value in the sample, which may not be the maximum
-// value ever to be part of the sample.
-func (s *ExpDecaySample) Max() int64 {
-	return SampleMax(s.Values())
-}
-
-// Mean returns the mean of the values in the sample.
-func (s *ExpDecaySample) Mean() float64 {
-	return SampleMean(s.Values())
-}
-
-// Min returns the minimum value in the sample, which may not be the minimum
-// value ever to be part of the sample.
-func (s *ExpDecaySample) Min() int64 {
-	return SampleMin(s.Values())
-}
-
-// Percentile returns an arbitrary percentile of values in the sample.
-func (s *ExpDecaySample) Percentile(p float64) float64 {
-	return SamplePercentile(s.Values(), p)
-}
-
-// Percentiles returns a slice of arbitrary percentiles of values in the
-// sample.
-func (s *ExpDecaySample) Percentiles(ps []float64) []float64 {
-	return SamplePercentiles(s.Values(), ps)
-}
-
-// Size returns the size of the sample, which is at most the reservoir size.
-func (s *ExpDecaySample) Size() int {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return s.values.Size()
-}
-
 // Snapshot returns a read-only copy of the sample.
-func (s *ExpDecaySample) Snapshot() Sample {
+func (s *ExpDecaySample) Snapshot() SampleSnapshot {
 	s.mutex.Lock()
 	defer s.mutex.Unlock()
-	vals := s.values.Values()
-	values := make([]int64, len(vals))
-	for i, v := range vals {
-		values[i] = v.v
-	}
-	return &SampleSnapshot{
-		count:  s.count,
-		values: values,
+	var (
+		samples       = s.values.Values()
+		values        = make([]int64, len(samples))
+		max     int64 = math.MinInt64
+		min     int64 = math.MaxInt64
+		sum     int64
+	)
+	for i, item := range samples {
+		v := item.v
+		values[i] = v
+		sum += v
+		if v > max {
+			max = v
+		}
+		if v < min {
+			min = v
+		}
 	}
-}
-
-// StdDev returns the standard deviation of the values in the sample.
-func (s *ExpDecaySample) StdDev() float64 {
-	return SampleStdDev(s.Values())
-}
-
-// Sum returns the sum of the values in the sample.
-func (s *ExpDecaySample) Sum() int64 {
-	return SampleSum(s.Values())
+	return newSampleSnapshotPrecalculated(s.count, values, min, max, sum)
 }
 
 // Update samples a new value.
@@ -150,23 +109,6 @@ func (s *ExpDecaySample) Update(v int64) {
 	s.update(time.Now(), v)
 }
 
-// Values returns a copy of the values in the sample.
-func (s *ExpDecaySample) Values() []int64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	vals := s.values.Values()
-	values := make([]int64, len(vals))
-	for i, v := range vals {
-		values[i] = v.v
-	}
-	return values
-}
-
-// Variance returns the variance of the values in the sample.
-func (s *ExpDecaySample) Variance() float64 {
-	return SampleVariance(s.Values())
-}
-
 // update samples a new value at a particular timestamp.  This is a method all
 // its own to facilitate testing.
 func (s *ExpDecaySample) update(t time.Time, v int64) {
@@ -202,207 +144,160 @@ func (s *ExpDecaySample) update(t time.Time, v int64) {
 // NilSample is a no-op Sample.
 type NilSample struct{}
 
-// Clear is a no-op.
-func (NilSample) Clear() {}
-
-// Count is a no-op.
-func (NilSample) Count() int64 { return 0 }
-
-// Max is a no-op.
-func (NilSample) Max() int64 { return 0 }
-
-// Mean is a no-op.
-func (NilSample) Mean() float64 { return 0.0 }
-
-// Min is a no-op.
-func (NilSample) Min() int64 { return 0 }
-
-// Percentile is a no-op.
-func (NilSample) Percentile(p float64) float64 { return 0.0 }
-
-// Percentiles is a no-op.
-func (NilSample) Percentiles(ps []float64) []float64 {
-	return make([]float64, len(ps))
-}
-
-// Size is a no-op.
-func (NilSample) Size() int { return 0 }
-
-// Sample is a no-op.
-func (NilSample) Snapshot() Sample { return NilSample{} }
-
-// StdDev is a no-op.
-func (NilSample) StdDev() float64 { return 0.0 }
-
-// Sum is a no-op.
-func (NilSample) Sum() int64 { return 0 }
-
-// Update is a no-op.
-func (NilSample) Update(v int64) {}
-
-// Values is a no-op.
-func (NilSample) Values() []int64 { return []int64{} }
-
-// Variance is a no-op.
-func (NilSample) Variance() float64 { return 0.0 }
-
-// SampleMax returns the maximum value of the slice of int64.
-func SampleMax(values []int64) int64 {
-	if len(values) == 0 {
-		return 0
-	}
-	var max int64 = math.MinInt64
-	for _, v := range values {
-		if max < v {
-			max = v
-		}
-	}
-	return max
-}
-
-// SampleMean returns the mean value of the slice of int64.
-func SampleMean(values []int64) float64 {
-	if len(values) == 0 {
-		return 0.0
-	}
-	return float64(SampleSum(values)) / float64(len(values))
-}
-
-// SampleMin returns the minimum value of the slice of int64.
-func SampleMin(values []int64) int64 {
-	if len(values) == 0 {
-		return 0
-	}
-	var min int64 = math.MaxInt64
-	for _, v := range values {
-		if min > v {
-			min = v
-		}
-	}
-	return min
-}
+func (NilSample) Clear()                   {}
+func (NilSample) Snapshot() SampleSnapshot { return (*emptySnapshot)(nil) }
+func (NilSample) Update(v int64)           {}
 
 // SamplePercentiles returns an arbitrary percentile of the slice of int64.
 func SamplePercentile(values []int64, p float64) float64 {
-	return SamplePercentiles(values, []float64{p})[0]
+	return CalculatePercentiles(values, []float64{p})[0]
 }
 
-// SamplePercentiles returns a slice of arbitrary percentiles of the slice of
-// int64.
-func SamplePercentiles(values []int64, ps []float64) []float64 {
+// CalculatePercentiles returns a slice of arbitrary percentiles of the slice of
+// int64. This method returns interpolated results, so e.g if there are only two
+// values, [0, 10], a 50% percentile will land between them.
+//
+// Note: As a side-effect, this method will also sort the slice of values.
+// Note2: The input format for percentiles is NOT percent! To express 50%, use 0.5, not 50.
+func CalculatePercentiles(values []int64, ps []float64) []float64 {
 	scores := make([]float64, len(ps))
 	size := len(values)
-	if size > 0 {
-		slices.Sort(values)
-		for i, p := range ps {
-			pos := p * float64(size+1)
-			if pos < 1.0 {
-				scores[i] = float64(values[0])
-			} else if pos >= float64(size) {
-				scores[i] = float64(values[size-1])
-			} else {
-				lower := float64(values[int(pos)-1])
-				upper := float64(values[int(pos)])
-				scores[i] = lower + (pos-math.Floor(pos))*(upper-lower)
-			}
+	if size == 0 {
+		return scores
+	}
+	slices.Sort(values)
+	for i, p := range ps {
+		pos := p * float64(size+1)
+
+		if pos < 1.0 {
+			scores[i] = float64(values[0])
+		} else if pos >= float64(size) {
+			scores[i] = float64(values[size-1])
+		} else {
+			lower := float64(values[int(pos)-1])
+			upper := float64(values[int(pos)])
+			scores[i] = lower + (pos-math.Floor(pos))*(upper-lower)
 		}
 	}
 	return scores
 }
 
-// SampleSnapshot is a read-only copy of another Sample.
-type SampleSnapshot struct {
+// sampleSnapshot is a read-only copy of another Sample.
+type sampleSnapshot struct {
 	count  int64
 	values []int64
+
+	max      int64
+	min      int64
+	mean     float64
+	sum      int64
+	variance float64
 }
 
-func NewSampleSnapshot(count int64, values []int64) *SampleSnapshot {
-	return &SampleSnapshot{
+// newSampleSnapshotPrecalculated creates a read-only sampleSnapShot, using
+// precalculated sums to avoid iterating the values
+func newSampleSnapshotPrecalculated(count int64, values []int64, min, max, sum int64) *sampleSnapshot {
+	if len(values) == 0 {
+		return &sampleSnapshot{
+			count:  count,
+			values: values,
+		}
+	}
+	return &sampleSnapshot{
 		count:  count,
 		values: values,
+		max:    max,
+		min:    min,
+		mean:   float64(sum) / float64(len(values)),
+		sum:    sum,
 	}
 }
 
-// Clear panics.
-func (*SampleSnapshot) Clear() {
-	panic("Clear called on a SampleSnapshot")
+// newSampleSnapshot creates a read-only sampleSnapShot, and calculates some
+// numbers.
+func newSampleSnapshot(count int64, values []int64) *sampleSnapshot {
+	var (
+		max int64 = math.MinInt64
+		min int64 = math.MaxInt64
+		sum int64
+	)
+	for _, v := range values {
+		sum += v
+		if v > max {
+			max = v
+		}
+		if v < min {
+			min = v
+		}
+	}
+	return newSampleSnapshotPrecalculated(count, values, min, max, sum)
 }
 
 // Count returns the count of inputs at the time the snapshot was taken.
-func (s *SampleSnapshot) Count() int64 { return s.count }
+func (s *sampleSnapshot) Count() int64 { return s.count }
 
 // Max returns the maximal value at the time the snapshot was taken.
-func (s *SampleSnapshot) Max() int64 { return SampleMax(s.values) }
+func (s *sampleSnapshot) Max() int64 { return s.max }
 
 // Mean returns the mean value at the time the snapshot was taken.
-func (s *SampleSnapshot) Mean() float64 { return SampleMean(s.values) }
+func (s *sampleSnapshot) Mean() float64 { return s.mean }
 
 // Min returns the minimal value at the time the snapshot was taken.
-func (s *SampleSnapshot) Min() int64 { return SampleMin(s.values) }
+func (s *sampleSnapshot) Min() int64 { return s.min }
 
 // Percentile returns an arbitrary percentile of values at the time the
 // snapshot was taken.
-func (s *SampleSnapshot) Percentile(p float64) float64 {
+func (s *sampleSnapshot) Percentile(p float64) float64 {
 	return SamplePercentile(s.values, p)
 }
 
 // Percentiles returns a slice of arbitrary percentiles of values at the time
 // the snapshot was taken.
-func (s *SampleSnapshot) Percentiles(ps []float64) []float64 {
-	return SamplePercentiles(s.values, ps)
+func (s *sampleSnapshot) Percentiles(ps []float64) []float64 {
+	return CalculatePercentiles(s.values, ps)
 }
 
 // Size returns the size of the sample at the time the snapshot was taken.
-func (s *SampleSnapshot) Size() int { return len(s.values) }
+func (s *sampleSnapshot) Size() int { return len(s.values) }
 
 // Snapshot returns the snapshot.
-func (s *SampleSnapshot) Snapshot() Sample { return s }
+func (s *sampleSnapshot) Snapshot() SampleSnapshot { return s }
 
 // StdDev returns the standard deviation of values at the time the snapshot was
 // taken.
-func (s *SampleSnapshot) StdDev() float64 { return SampleStdDev(s.values) }
+func (s *sampleSnapshot) StdDev() float64 {
+	if s.variance == 0.0 {
+		s.variance = SampleVariance(s.mean, s.values)
+	}
+	return math.Sqrt(s.variance)
+}
 
 // Sum returns the sum of values at the time the snapshot was taken.
-func (s *SampleSnapshot) Sum() int64 { return SampleSum(s.values) }
-
-// Update panics.
-func (*SampleSnapshot) Update(int64) {
-	panic("Update called on a SampleSnapshot")
-}
+func (s *sampleSnapshot) Sum() int64 { return s.sum }
 
 // Values returns a copy of the values in the sample.
-func (s *SampleSnapshot) Values() []int64 {
+func (s *sampleSnapshot) Values() []int64 {
 	values := make([]int64, len(s.values))
 	copy(values, s.values)
 	return values
 }
 
 // Variance returns the variance of values at the time the snapshot was taken.
-func (s *SampleSnapshot) Variance() float64 { return SampleVariance(s.values) }
-
-// SampleStdDev returns the standard deviation of the slice of int64.
-func SampleStdDev(values []int64) float64 {
-	return math.Sqrt(SampleVariance(values))
-}
-
-// SampleSum returns the sum of the slice of int64.
-func SampleSum(values []int64) int64 {
-	var sum int64
-	for _, v := range values {
-		sum += v
+func (s *sampleSnapshot) Variance() float64 {
+	if s.variance == 0.0 {
+		s.variance = SampleVariance(s.mean, s.values)
 	}
-	return sum
+	return s.variance
 }
 
 // SampleVariance returns the variance of the slice of int64.
-func SampleVariance(values []int64) float64 {
+func SampleVariance(mean float64, values []int64) float64 {
 	if len(values) == 0 {
 		return 0.0
 	}
-	m := SampleMean(values)
 	var sum float64
 	for _, v := range values {
-		d := float64(v) - m
+		d := float64(v) - mean
 		sum += d * d
 	}
 	return sum / float64(len(values))
@@ -445,83 +340,14 @@ func (s *UniformSample) Clear() {
 	s.values = make([]int64, 0, s.reservoirSize)
 }
 
-// Count returns the number of samples recorded, which may exceed the
-// reservoir size.
-func (s *UniformSample) Count() int64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return s.count
-}
-
-// Max returns the maximum value in the sample, which may not be the maximum
-// value ever to be part of the sample.
-func (s *UniformSample) Max() int64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return SampleMax(s.values)
-}
-
-// Mean returns the mean of the values in the sample.
-func (s *UniformSample) Mean() float64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return SampleMean(s.values)
-}
-
-// Min returns the minimum value in the sample, which may not be the minimum
-// value ever to be part of the sample.
-func (s *UniformSample) Min() int64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return SampleMin(s.values)
-}
-
-// Percentile returns an arbitrary percentile of values in the sample.
-func (s *UniformSample) Percentile(p float64) float64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return SamplePercentile(s.values, p)
-}
-
-// Percentiles returns a slice of arbitrary percentiles of values in the
-// sample.
-func (s *UniformSample) Percentiles(ps []float64) []float64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return SamplePercentiles(s.values, ps)
-}
-
-// Size returns the size of the sample, which is at most the reservoir size.
-func (s *UniformSample) Size() int {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return len(s.values)
-}
-
 // Snapshot returns a read-only copy of the sample.
-func (s *UniformSample) Snapshot() Sample {
+func (s *UniformSample) Snapshot() SampleSnapshot {
 	s.mutex.Lock()
-	defer s.mutex.Unlock()
 	values := make([]int64, len(s.values))
 	copy(values, s.values)
-	return &SampleSnapshot{
-		count:  s.count,
-		values: values,
-	}
-}
-
-// StdDev returns the standard deviation of the values in the sample.
-func (s *UniformSample) StdDev() float64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return SampleStdDev(s.values)
-}
-
-// Sum returns the sum of the values in the sample.
-func (s *UniformSample) Sum() int64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return SampleSum(s.values)
+	count := s.count
+	s.mutex.Unlock()
+	return newSampleSnapshot(count, values)
 }
 
 // Update samples a new value.
@@ -544,22 +370,6 @@ func (s *UniformSample) Update(v int64) {
 	}
 }
 
-// Values returns a copy of the values in the sample.
-func (s *UniformSample) Values() []int64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	values := make([]int64, len(s.values))
-	copy(values, s.values)
-	return values
-}
-
-// Variance returns the variance of the values in the sample.
-func (s *UniformSample) Variance() float64 {
-	s.mutex.Lock()
-	defer s.mutex.Unlock()
-	return SampleVariance(s.values)
-}
-
 // expDecaySample represents an individual sample in a heap.
 type expDecaySample struct {
 	k float64
diff --git a/metrics/sample_test.go b/metrics/sample_test.go
index 3ae128d56f..7967357055 100644
--- a/metrics/sample_test.go
+++ b/metrics/sample_test.go
@@ -8,28 +8,36 @@ import (
 	"time"
 )
 
+const epsilonPercentile = .00000000001
+
 // Benchmark{Compute,Copy}{1000,1000000} demonstrate that, even for relatively
 // expensive computations like Variance, the cost of copying the Sample, as
 // approximated by a make and copy, is much greater than the cost of the
 // computation for small samples and only slightly less for large samples.
 func BenchmarkCompute1000(b *testing.B) {
 	s := make([]int64, 1000)
+	var sum int64
 	for i := 0; i < len(s); i++ {
 		s[i] = int64(i)
+		sum += int64(i)
 	}
+	mean := float64(sum) / float64(len(s))
 	b.ResetTimer()
 	for i := 0; i < b.N; i++ {
-		SampleVariance(s)
+		SampleVariance(mean, s)
 	}
 }
 func BenchmarkCompute1000000(b *testing.B) {
 	s := make([]int64, 1000000)
+	var sum int64
 	for i := 0; i < len(s); i++ {
 		s[i] = int64(i)
+		sum += int64(i)
 	}
+	mean := float64(sum) / float64(len(s))
 	b.ResetTimer()
 	for i := 0; i < b.N; i++ {
-		SampleVariance(s)
+		SampleVariance(mean, s)
 	}
 }
 func BenchmarkCopy1000(b *testing.B) {
@@ -79,65 +87,42 @@ func BenchmarkUniformSample1028(b *testing.B) {
 	benchmarkSample(b, NewUniformSample(1028))
 }
 
-func TestExpDecaySample10(t *testing.T) {
-	s := NewExpDecaySample(100, 0.99)
-	for i := 0; i < 10; i++ {
-		s.Update(int64(i))
-	}
-	if size := s.Count(); size != 10 {
-		t.Errorf("s.Count(): 10 != %v\n", size)
-	}
-	if size := s.Size(); size != 10 {
-		t.Errorf("s.Size(): 10 != %v\n", size)
-	}
-	if l := len(s.Values()); l != 10 {
-		t.Errorf("len(s.Values()): 10 != %v\n", l)
-	}
-	for _, v := range s.Values() {
-		if v > 10 || v < 0 {
-			t.Errorf("out of range [0, 10): %v\n", v)
-		}
+func min(a, b int) int {
+	if a < b {
+		return a
 	}
+	return b
 }
 
-func TestExpDecaySample100(t *testing.T) {
-	s := NewExpDecaySample(1000, 0.01)
-	for i := 0; i < 100; i++ {
-		s.Update(int64(i))
-	}
-	if size := s.Count(); size != 100 {
-		t.Errorf("s.Count(): 100 != %v\n", size)
-	}
-	if size := s.Size(); size != 100 {
-		t.Errorf("s.Size(): 100 != %v\n", size)
-	}
-	if l := len(s.Values()); l != 100 {
-		t.Errorf("len(s.Values()): 100 != %v\n", l)
-	}
-	for _, v := range s.Values() {
-		if v > 100 || v < 0 {
-			t.Errorf("out of range [0, 100): %v\n", v)
+func TestExpDecaySample(t *testing.T) {
+	for _, tc := range []struct {
+		reservoirSize int
+		alpha         float64
+		updates       int
+	}{
+		{100, 0.99, 10},
+		{1000, 0.01, 100},
+		{100, 0.99, 1000},
+	} {
+		sample := NewExpDecaySample(tc.reservoirSize, tc.alpha)
+		for i := 0; i < tc.updates; i++ {
+			sample.Update(int64(i))
 		}
-	}
-}
-
-func TestExpDecaySample1000(t *testing.T) {
-	s := NewExpDecaySample(100, 0.99)
-	for i := 0; i < 1000; i++ {
-		s.Update(int64(i))
-	}
-	if size := s.Count(); size != 1000 {
-		t.Errorf("s.Count(): 1000 != %v\n", size)
-	}
-	if size := s.Size(); size != 100 {
-		t.Errorf("s.Size(): 100 != %v\n", size)
-	}
-	if l := len(s.Values()); l != 100 {
-		t.Errorf("len(s.Values()): 100 != %v\n", l)
-	}
-	for _, v := range s.Values() {
-		if v > 1000 || v < 0 {
-			t.Errorf("out of range [0, 1000): %v\n", v)
+		snap := sample.Snapshot()
+		if have, want := int(snap.Count()), tc.updates; have != want {
+			t.Errorf("have %d want %d", have, want)
+		}
+		if have, want := snap.Size(), min(tc.updates, tc.reservoirSize); have != want {
+			t.Errorf("have %d want %d", have, want)
+		}
+		values := snap.(*sampleSnapshot).values
+		if have, want := len(values), min(tc.updates, tc.reservoirSize); have != want {
+			t.Errorf("have %d want %d", have, want)
+		}
+		for _, v := range values {
+			if v > int64(tc.updates) || v < 0 {
+				t.Errorf("out of range [0, %d): %v", tc.updates, v)
+			}
 		}
 	}
 }
@@ -147,15 +132,16 @@ func TestExpDecaySample1000(t *testing.T) {
 // The priority becomes +Inf quickly after starting if this is done,
 // effectively freezing the set of samples until a rescale step happens.
 func TestExpDecaySampleNanosecondRegression(t *testing.T) {
-	s := NewExpDecaySample(100, 0.99)
+	sw := NewExpDecaySample(100, 0.99)
 	for i := 0; i < 100; i++ {
-		s.Update(10)
+		sw.Update(10)
 	}
 	time.Sleep(1 * time.Millisecond)
 	for i := 0; i < 100; i++ {
-		s.Update(20)
+		sw.Update(20)
 	}
-	v := s.Values()
+	s := sw.Snapshot()
+	v := s.(*sampleSnapshot).values
 	avg := float64(0)
 	for i := 0; i < len(v); i++ {
 		avg += float64(v[i])
@@ -194,24 +180,27 @@ func TestExpDecaySampleStatistics(t *testing.T) {
 	for i := 1; i <= 10000; i++ {
 		s.(*ExpDecaySample).update(now.Add(time.Duration(i)), int64(i))
 	}
-	testExpDecaySampleStatistics(t, s)
+	testExpDecaySampleStatistics(t, s.Snapshot())
 }
 
 func TestUniformSample(t *testing.T) {
-	s := NewUniformSample(100)
+	sw := NewUniformSample(100)
 	for i := 0; i < 1000; i++ {
-		s.Update(int64(i))
+		sw.Update(int64(i))
 	}
+	s := sw.Snapshot()
 	if size := s.Count(); size != 1000 {
 		t.Errorf("s.Count(): 1000 != %v\n", size)
 	}
 	if size := s.Size(); size != 100 {
 		t.Errorf("s.Size(): 100 != %v\n", size)
 	}
-	if l := len(s.Values()); l != 100 {
+	values := s.(*sampleSnapshot).values
+
+	if l := len(values); l != 100 {
 		t.Errorf("len(s.Values()): 100 != %v\n", l)
 	}
-	for _, v := range s.Values() {
+	for _, v := range values {
 		if v > 1000 || v < 0 {
 			t.Errorf("out of range [0, 100): %v\n", v)
 		}
@@ -219,12 +208,13 @@ func TestUniformSample(t *testing.T) {
 }
 
 func TestUniformSampleIncludesTail(t *testing.T) {
-	s := NewUniformSample(100)
+	sw := NewUniformSample(100)
 	max := 100
 	for i := 0; i < max; i++ {
-		s.Update(int64(i))
+		sw.Update(int64(i))
 	}
-	v := s.Values()
+	s := sw.Snapshot()
+	v := s.(*sampleSnapshot).values
 	sum := 0
 	exp := (max - 1) * max / 2
 	for i := 0; i < len(v); i++ {
@@ -250,7 +240,7 @@ func TestUniformSampleStatistics(t *testing.T) {
 	for i := 1; i <= 10000; i++ {
 		s.Update(int64(i))
 	}
-	testUniformSampleStatistics(t, s)
+	testUniformSampleStatistics(t, s.Snapshot())
 }
 
 func benchmarkSample(b *testing.B, s Sample) {
@@ -267,7 +257,7 @@ func benchmarkSample(b *testing.B, s Sample) {
 	b.Logf("GC cost: %d ns/op", int(memStats.PauseTotalNs-pauseTotalNs)/b.N)
 }
 
-func testExpDecaySampleStatistics(t *testing.T, s Sample) {
+func testExpDecaySampleStatistics(t *testing.T, s SampleSnapshot) {
 	if count := s.Count(); count != 10000 {
 		t.Errorf("s.Count(): 10000 != %v\n", count)
 	}
@@ -295,7 +285,7 @@ func testExpDecaySampleStatistics(t *testing.T, s Sample) {
 	}
 }
 
-func testUniformSampleStatistics(t *testing.T, s Sample) {
+func testUniformSampleStatistics(t *testing.T, s SampleSnapshot) {
 	if count := s.Count(); count != 10000 {
 		t.Errorf("s.Count(): 10000 != %v\n", count)
 	}
@@ -349,8 +339,22 @@ func TestUniformSampleConcurrentUpdateCount(t *testing.T) {
 		}
 	}()
 	for i := 0; i < 1000; i++ {
-		s.Count()
+		s.Snapshot().Count()
 		time.Sleep(5 * time.Millisecond)
 	}
 	quit <- struct{}{}
 }
+
+func BenchmarkCalculatePercentiles(b *testing.B) {
+	pss := []float64{0.5, 0.75, 0.95, 0.99, 0.999, 0.9999}
+	var vals []int64
+	for i := 0; i < 1000; i++ {
+		vals = append(vals, int64(rand.Int31()))
+	}
+	v := make([]int64, len(vals))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		copy(v, vals)
+		_ = CalculatePercentiles(v, pss)
+	}
+}
diff --git a/metrics/syslog.go b/metrics/syslog.go
index 76c8490567..fd856d6973 100644
--- a/metrics/syslog.go
+++ b/metrics/syslog.go
@@ -16,15 +16,15 @@ func Syslog(r Registry, d time.Duration, w *syslog.Writer) {
 		r.Each(func(name string, i interface{}) {
 			switch metric := i.(type) {
 			case Counter:
-				w.Info(fmt.Sprintf("counter %s: count: %d", name, metric.Count()))
+				w.Info(fmt.Sprintf("counter %s: count: %d", name, metric.Snapshot().Count()))
 			case CounterFloat64:
-				w.Info(fmt.Sprintf("counter %s: count: %f", name, metric.Count()))
+				w.Info(fmt.Sprintf("counter %s: count: %f", name, metric.Snapshot().Count()))
 			case Gauge:
-				w.Info(fmt.Sprintf("gauge %s: value: %d", name, metric.Value()))
+				w.Info(fmt.Sprintf("gauge %s: value: %d", name, metric.Snapshot().Value()))
 			case GaugeFloat64:
-				w.Info(fmt.Sprintf("gauge %s: value: %f", name, metric.Value()))
+				w.Info(fmt.Sprintf("gauge %s: value: %f", name, metric.Snapshot().Value()))
 			case GaugeInfo:
-				w.Info(fmt.Sprintf("gauge %s: value: %s", name, metric.Value()))
+				w.Info(fmt.Sprintf("gauge %s: value: %s", name, metric.Snapshot().Value()))
 			case Healthcheck:
 				metric.Check()
 				w.Info(fmt.Sprintf("healthcheck %s: error: %v", name, metric.Error()))
diff --git a/metrics/testdata/opentsb.want b/metrics/testdata/opentsb.want
index c8e40a5250..43fe1b2ac2 100644
--- a/metrics/testdata/opentsb.want
+++ b/metrics/testdata/opentsb.want
@@ -1,4 +1,4 @@
-put pre.elite.count 978307200 0 host=hal9000
+put pre.elite.count 978307200 1337 host=hal9000
 put pre.elite.one-minute 978307200 0.00 host=hal9000
 put pre.elite.five-minute 978307200 0.00 host=hal9000
 put pre.elite.fifteen-minute 978307200 0.00 host=hal9000
diff --git a/metrics/timer.go b/metrics/timer.go
index 2e1a9be472..576ad8aa3e 100644
--- a/metrics/timer.go
+++ b/metrics/timer.go
@@ -5,26 +5,18 @@ import (
 	"time"
 )
 
+type TimerSnapshot interface {
+	HistogramSnapshot
+	MeterSnapshot
+}
+
 // Timers capture the duration and rate of events.
 type Timer interface {
-	Count() int64
-	Max() int64
-	Mean() float64
-	Min() int64
-	Percentile(float64) float64
-	Percentiles([]float64) []float64
-	Rate1() float64
-	Rate5() float64
-	Rate15() float64
-	RateMean() float64
-	Snapshot() Timer
-	StdDev() float64
+	Snapshot() TimerSnapshot
 	Stop()
-	Sum() int64
 	Time(func())
-	Update(time.Duration)
 	UpdateSince(time.Time)
-	Variance() float64
+	Update(time.Duration)
 }
 
 // GetOrRegisterTimer returns an existing Timer or constructs and registers a
@@ -78,61 +70,11 @@ func NewTimer() Timer {
 // NilTimer is a no-op Timer.
 type NilTimer struct{}
 
-// Count is a no-op.
-func (NilTimer) Count() int64 { return 0 }
-
-// Max is a no-op.
-func (NilTimer) Max() int64 { return 0 }
-
-// Mean is a no-op.
-func (NilTimer) Mean() float64 { return 0.0 }
-
-// Min is a no-op.
-func (NilTimer) Min() int64 { return 0 }
-
-// Percentile is a no-op.
-func (NilTimer) Percentile(p float64) float64 { return 0.0 }
-
-// Percentiles is a no-op.
-func (NilTimer) Percentiles(ps []float64) []float64 {
-	return make([]float64, len(ps))
-}
-
-// Rate1 is a no-op.
-func (NilTimer) Rate1() float64 { return 0.0 }
-
-// Rate5 is a no-op.
-func (NilTimer) Rate5() float64 { return 0.0 }
-
-// Rate15 is a no-op.
-func (NilTimer) Rate15() float64 { return 0.0 }
-
-// RateMean is a no-op.
-func (NilTimer) RateMean() float64 { return 0.0 }
-
-// Snapshot is a no-op.
-func (NilTimer) Snapshot() Timer { return NilTimer{} }
-
-// StdDev is a no-op.
-func (NilTimer) StdDev() float64 { return 0.0 }
-
-// Stop is a no-op.
-func (NilTimer) Stop() {}
-
-// Sum is a no-op.
-func (NilTimer) Sum() int64 { return 0 }
-
-// Time is a no-op.
-func (NilTimer) Time(f func()) { f() }
-
-// Update is a no-op.
-func (NilTimer) Update(time.Duration) {}
-
-// UpdateSince is a no-op.
-func (NilTimer) UpdateSince(time.Time) {}
-
-// Variance is a no-op.
-func (NilTimer) Variance() float64 { return 0.0 }
+func (NilTimer) Snapshot() TimerSnapshot { return (*emptySnapshot)(nil) }
+func (NilTimer) Stop()                   {}
+func (NilTimer) Time(f func())           { f() }
+func (NilTimer) Update(time.Duration)    {}
+func (NilTimer) UpdateSince(time.Time)   {}
 
 // StandardTimer is the standard implementation of a Timer and uses a Histogram
 // and Meter.
@@ -142,82 +84,21 @@ type StandardTimer struct {
 	mutex     sync.Mutex
 }
 
-// Count returns the number of events recorded.
-func (t *StandardTimer) Count() int64 {
-	return t.histogram.Count()
-}
-
-// Max returns the maximum value in the sample.
-func (t *StandardTimer) Max() int64 {
-	return t.histogram.Max()
-}
-
-// Mean returns the mean of the values in the sample.
-func (t *StandardTimer) Mean() float64 {
-	return t.histogram.Mean()
-}
-
-// Min returns the minimum value in the sample.
-func (t *StandardTimer) Min() int64 {
-	return t.histogram.Min()
-}
-
-// Percentile returns an arbitrary percentile of the values in the sample.
-func (t *StandardTimer) Percentile(p float64) float64 {
-	return t.histogram.Percentile(p)
-}
-
-// Percentiles returns a slice of arbitrary percentiles of the values in the
-// sample.
-func (t *StandardTimer) Percentiles(ps []float64) []float64 {
-	return t.histogram.Percentiles(ps)
-}
-
-// Rate1 returns the one-minute moving average rate of events per second.
-func (t *StandardTimer) Rate1() float64 {
-	return t.meter.Rate1()
-}
-
-// Rate5 returns the five-minute moving average rate of events per second.
-func (t *StandardTimer) Rate5() float64 {
-	return t.meter.Rate5()
-}
-
-// Rate15 returns the fifteen-minute moving average rate of events per second.
-func (t *StandardTimer) Rate15() float64 {
-	return t.meter.Rate15()
-}
-
-// RateMean returns the meter's mean rate of events per second.
-func (t *StandardTimer) RateMean() float64 {
-	return t.meter.RateMean()
-}
-
 // Snapshot returns a read-only copy of the timer.
-func (t *StandardTimer) Snapshot() Timer {
+func (t *StandardTimer) Snapshot() TimerSnapshot {
 	t.mutex.Lock()
 	defer t.mutex.Unlock()
-	return &TimerSnapshot{
-		histogram: t.histogram.Snapshot().(*HistogramSnapshot),
-		meter:     t.meter.Snapshot().(*MeterSnapshot),
+	return &timerSnapshot{
+		histogram: t.histogram.Snapshot(),
+		meter:     t.meter.Snapshot(),
 	}
 }
 
-// StdDev returns the standard deviation of the values in the sample.
-func (t *StandardTimer) StdDev() float64 {
-	return t.histogram.StdDev()
-}
-
 // Stop stops the meter.
 func (t *StandardTimer) Stop() {
 	t.meter.Stop()
 }
 
-// Sum returns the sum in the sample.
-func (t *StandardTimer) Sum() int64 {
-	return t.histogram.Sum()
-}
-
 // Record the duration of the execution of the given function.
 func (t *StandardTimer) Time(f func()) {
 	ts := time.Now()
@@ -241,86 +122,63 @@ func (t *StandardTimer) UpdateSince(ts time.Time) {
 	t.meter.Mark(1)
 }
 
-// Variance returns the variance of the values in the sample.
-func (t *StandardTimer) Variance() float64 {
-	return t.histogram.Variance()
-}
-
-// TimerSnapshot is a read-only copy of another Timer.
-type TimerSnapshot struct {
-	histogram *HistogramSnapshot
-	meter     *MeterSnapshot
+// timerSnapshot is a read-only copy of another Timer.
+type timerSnapshot struct {
+	histogram HistogramSnapshot
+	meter     MeterSnapshot
 }
 
 // Count returns the number of events recorded at the time the snapshot was
 // taken.
-func (t *TimerSnapshot) Count() int64 { return t.histogram.Count() }
+func (t *timerSnapshot) Count() int64 { return t.histogram.Count() }
 
 // Max returns the maximum value at the time the snapshot was taken.
-func (t *TimerSnapshot) Max() int64 { return t.histogram.Max() }
+func (t *timerSnapshot) Max() int64 { return t.histogram.Max() }
+
+// Size returns the size of the sample at the time the snapshot was taken.
+func (t *timerSnapshot) Size() int { return t.histogram.Size() }
 
 // Mean returns the mean value at the time the snapshot was taken.
-func (t *TimerSnapshot) Mean() float64 { return t.histogram.Mean() }
+func (t *timerSnapshot) Mean() float64 { return t.histogram.Mean() }
 
 // Min returns the minimum value at the time the snapshot was taken.
-func (t *TimerSnapshot) Min() int64 { return t.histogram.Min() }
+func (t *timerSnapshot) Min() int64 { return t.histogram.Min() }
 
 // Percentile returns an arbitrary percentile of sampled values at the time the
 // snapshot was taken.
-func (t *TimerSnapshot) Percentile(p float64) float64 {
+func (t *timerSnapshot) Percentile(p float64) float64 {
 	return t.histogram.Percentile(p)
 }
 
 // Percentiles returns a slice of arbitrary percentiles of sampled values at
 // the time the snapshot was taken.
-func (t *TimerSnapshot) Percentiles(ps []float64) []float64 {
+func (t *timerSnapshot) Percentiles(ps []float64) []float64 {
 	return t.histogram.Percentiles(ps)
 }
 
 // Rate1 returns the one-minute moving average rate of events per second at the
 // time the snapshot was taken.
-func (t *TimerSnapshot) Rate1() float64 { return t.meter.Rate1() }
+func (t *timerSnapshot) Rate1() float64 { return t.meter.Rate1() }
 
 // Rate5 returns the five-minute moving average rate of events per second at
 // the time the snapshot was taken.
-func (t *TimerSnapshot) Rate5() float64 { return t.meter.Rate5() }
+func (t *timerSnapshot) Rate5() float64 { return t.meter.Rate5() }
 
 // Rate15 returns the fifteen-minute moving average rate of events per second
 // at the time the snapshot was taken.
-func (t *TimerSnapshot) Rate15() float64 { return t.meter.Rate15() }
+func (t *timerSnapshot) Rate15() float64 { return t.meter.Rate15() }
 
 // RateMean returns the meter's mean rate of events per second at the time the
 // snapshot was taken.
-func (t *TimerSnapshot) RateMean() float64 { return t.meter.RateMean() }
-
-// Snapshot returns the snapshot.
-func (t *TimerSnapshot) Snapshot() Timer { return t }
+func (t *timerSnapshot) RateMean() float64 { return t.meter.RateMean() }
 
 // StdDev returns the standard deviation of the values at the time the snapshot
 // was taken.
-func (t *TimerSnapshot) StdDev() float64 { return t.histogram.StdDev() }
-
-// Stop is a no-op.
-func (t *TimerSnapshot) Stop() {}
+func (t *timerSnapshot) StdDev() float64 { return t.histogram.StdDev() }
 
 // Sum returns the sum at the time the snapshot was taken.
-func (t *TimerSnapshot) Sum() int64 { return t.histogram.Sum() }
-
-// Time panics.
-func (*TimerSnapshot) Time(func()) {
-	panic("Time called on a TimerSnapshot")
-}
-
-// Update panics.
-func (*TimerSnapshot) Update(time.Duration) {
-	panic("Update called on a TimerSnapshot")
-}
-
-// UpdateSince panics.
-func (*TimerSnapshot) UpdateSince(time.Time) {
-	panic("UpdateSince called on a TimerSnapshot")
-}
+func (t *timerSnapshot) Sum() int64 { return t.histogram.Sum() }
 
 // Variance returns the variance of the values at the time the snapshot was
 // taken.
-func (t *TimerSnapshot) Variance() float64 { return t.histogram.Variance() }
+func (t *timerSnapshot) Variance() float64 { return t.histogram.Variance() }
diff --git a/metrics/timer_test.go b/metrics/timer_test.go
index 903e8e8d49..f10de16c9c 100644
--- a/metrics/timer_test.go
+++ b/metrics/timer_test.go
@@ -18,7 +18,7 @@ func BenchmarkTimer(b *testing.B) {
 func TestGetOrRegisterTimer(t *testing.T) {
 	r := NewRegistry()
 	NewRegisteredTimer("foo", r).Update(47)
-	if tm := GetOrRegisterTimer("foo", r); tm.Count() != 1 {
+	if tm := GetOrRegisterTimer("foo", r).Snapshot(); tm.Count() != 1 {
 		t.Fatal(tm)
 	}
 }
@@ -27,7 +27,7 @@ func TestTimerExtremes(t *testing.T) {
 	tm := NewTimer()
 	tm.Update(math.MaxInt64)
 	tm.Update(0)
-	if stdDev := tm.StdDev(); stdDev != 4.611686018427388e+18 {
+	if stdDev := tm.Snapshot().StdDev(); stdDev != 4.611686018427388e+18 {
 		t.Errorf("tm.StdDev(): 4.611686018427388e+18 != %v\n", stdDev)
 	}
 }
@@ -56,7 +56,7 @@ func TestTimerFunc(t *testing.T) {
 	})
 	var (
 		drift    = time.Millisecond * 2
-		measured = time.Duration(tm.Max())
+		measured = time.Duration(tm.Snapshot().Max())
 		ceil     = actualTime + drift
 		floor    = actualTime - drift
 	)
@@ -66,7 +66,7 @@ func TestTimerFunc(t *testing.T) {
 }
 
 func TestTimerZero(t *testing.T) {
-	tm := NewTimer()
+	tm := NewTimer().Snapshot()
 	if count := tm.Count(); count != 0 {
 		t.Errorf("tm.Count(): 0 != %v\n", count)
 	}
@@ -110,5 +110,5 @@ func ExampleGetOrRegisterTimer() {
 	m := "account.create.latency"
 	t := GetOrRegisterTimer(m, nil)
 	t.Update(47)
-	fmt.Println(t.Max()) // Output: 47
+	fmt.Println(t.Snapshot().Max()) // Output: 47
 }
diff --git a/metrics/writer.go b/metrics/writer.go
index ec2e4f8c6a..098da45c27 100644
--- a/metrics/writer.go
+++ b/metrics/writer.go
@@ -29,19 +29,19 @@ func WriteOnce(r Registry, w io.Writer) {
 		switch metric := namedMetric.m.(type) {
 		case Counter:
 			fmt.Fprintf(w, "counter %s\n", namedMetric.name)
-			fmt.Fprintf(w, "  count:       %9d\n", metric.Count())
+			fmt.Fprintf(w, "  count:       %9d\n", metric.Snapshot().Count())
 		case CounterFloat64:
 			fmt.Fprintf(w, "counter %s\n", namedMetric.name)
-			fmt.Fprintf(w, "  count:       %f\n", metric.Count())
+			fmt.Fprintf(w, "  count:       %f\n", metric.Snapshot().Count())
 		case Gauge:
 			fmt.Fprintf(w, "gauge %s\n", namedMetric.name)
-			fmt.Fprintf(w, "  value:       %9d\n", metric.Value())
+			fmt.Fprintf(w, "  value:       %9d\n", metric.Snapshot().Value())
 		case GaugeFloat64:
 			fmt.Fprintf(w, "gauge %s\n", namedMetric.name)
-			fmt.Fprintf(w, "  value:       %f\n", metric.Value())
+			fmt.Fprintf(w, "  value:       %f\n", metric.Snapshot().Value())
 		case GaugeInfo:
 			fmt.Fprintf(w, "gauge %s\n", namedMetric.name)
-			fmt.Fprintf(w, "  value:       %s\n", metric.Value().String())
+			fmt.Fprintf(w, "  value:       %s\n", metric.Snapshot().Value().String())
 		case Healthcheck:
 			metric.Check()
 			fmt.Fprintf(w, "healthcheck %s\n", namedMetric.name)