go-pulse/metrics/librato/librato.go
Martin Holst Swende 8b6cf128af
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.
2023-09-13 13:13:47 -04:00

255 lines
8.3 KiB
Go

package librato
import (
"fmt"
"log"
"math"
"regexp"
"time"
"github.com/ethereum/go-ethereum/metrics"
)
// a regexp for extracting the unit from time.Duration.String
var unitRegexp = regexp.MustCompile(`[^\\d]+$`)
// a helper that turns a time.Duration into librato display attributes for timer metrics
func translateTimerAttributes(d time.Duration) (attrs map[string]interface{}) {
attrs = make(map[string]interface{})
attrs[DisplayTransform] = fmt.Sprintf("x/%d", int64(d))
attrs[DisplayUnitsShort] = string(unitRegexp.Find([]byte(d.String())))
return
}
type Reporter struct {
Email, Token string
Namespace string
Source string
Interval time.Duration
Registry metrics.Registry
Percentiles []float64 // percentiles to report on histogram metrics
TimerAttributes map[string]interface{} // units in which timers will be displayed
intervalSec int64
}
func NewReporter(r metrics.Registry, d time.Duration, e string, t string, s string, p []float64, u time.Duration) *Reporter {
return &Reporter{e, t, "", s, d, r, p, translateTimerAttributes(u), int64(d / time.Second)}
}
func Librato(r metrics.Registry, d time.Duration, e string, t string, s string, p []float64, u time.Duration) {
NewReporter(r, d, e, t, s, p, u).Run()
}
func (rep *Reporter) Run() {
log.Printf("WARNING: This client has been DEPRECATED! It has been moved to https://github.com/mihasya/go-metrics-librato and will be removed from rcrowley/go-metrics on August 5th 2015")
ticker := time.NewTicker(rep.Interval)
defer ticker.Stop()
metricsApi := &LibratoClient{rep.Email, rep.Token}
for now := range ticker.C {
var metrics Batch
var err error
if metrics, err = rep.BuildRequest(now, rep.Registry); err != nil {
log.Printf("ERROR constructing librato request body %s", err)
continue
}
if err := metricsApi.PostMetrics(metrics); err != nil {
log.Printf("ERROR sending metrics to librato %s", err)
continue
}
}
}
// calculate sum of squares from data provided by metrics.Histogram
// see http://en.wikipedia.org/wiki/Standard_deviation#Rapid_calculation_methods
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.TimerSnapshot) float64 {
count := float64(t.Count())
sumSquared := math.Pow(count*t.Mean(), 2)
sumSquares := math.Pow(count*t.StdDev(), 2) + sumSquared/count
if math.IsNaN(sumSquares) {
return 0.0
}
return sumSquares
}
func (rep *Reporter) BuildRequest(now time.Time, r metrics.Registry) (snapshot Batch, err error) {
snapshot = Batch{
// coerce timestamps to a stepping fn so that they line up in Librato graphs
MeasureTime: (now.Unix() / rep.intervalSec) * rep.intervalSec,
Source: rep.Source,
}
snapshot.Gauges = make([]Measurement, 0)
snapshot.Counters = make([]Measurement, 0)
histogramGaugeCount := 1 + len(rep.Percentiles)
r.Each(func(name string, metric interface{}) {
if rep.Namespace != "" {
name = fmt.Sprintf("%s.%s", rep.Namespace, name)
}
measurement := Measurement{}
measurement[Period] = rep.Interval.Seconds()
switch m := metric.(type) {
case metrics.Counter:
ms := m.Snapshot()
if ms.Count() > 0 {
measurement[Name] = fmt.Sprintf("%s.%s", name, "count")
measurement[Value] = float64(ms.Count())
measurement[Attributes] = map[string]interface{}{
DisplayUnitsLong: Operations,
DisplayUnitsShort: OperationsShort,
DisplayMin: "0",
}
snapshot.Counters = append(snapshot.Counters, measurement)
}
case metrics.CounterFloat64:
if count := m.Snapshot().Count(); count > 0 {
measurement[Name] = fmt.Sprintf("%s.%s", name, "count")
measurement[Value] = count
measurement[Attributes] = map[string]interface{}{
DisplayUnitsLong: Operations,
DisplayUnitsShort: OperationsShort,
DisplayMin: "0",
}
snapshot.Counters = append(snapshot.Counters, measurement)
}
case metrics.Gauge:
measurement[Name] = name
measurement[Value] = float64(m.Snapshot().Value())
snapshot.Gauges = append(snapshot.Gauges, measurement)
case metrics.GaugeFloat64:
measurement[Name] = name
measurement[Value] = m.Snapshot().Value()
snapshot.Gauges = append(snapshot.Gauges, measurement)
case metrics.GaugeInfo:
measurement[Name] = name
measurement[Value] = m.Snapshot().Value()
snapshot.Gauges = append(snapshot.Gauges, measurement)
case metrics.Histogram:
ms := m.Snapshot()
if ms.Count() > 0 {
gauges := make([]Measurement, histogramGaugeCount)
measurement[Name] = fmt.Sprintf("%s.%s", name, "hist")
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: ms.Percentile(p),
Period: measurement[Period],
}
}
snapshot.Gauges = append(snapshot.Gauges, gauges...)
}
case metrics.Meter:
ms := m.Snapshot()
measurement[Name] = name
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: ms.Rate1(),
Period: int64(rep.Interval.Seconds()),
Attributes: map[string]interface{}{
DisplayUnitsLong: Operations,
DisplayUnitsShort: OperationsShort,
DisplayMin: "0",
},
},
Measurement{
Name: fmt.Sprintf("%s.%s", name, "5min"),
Value: ms.Rate5(),
Period: int64(rep.Interval.Seconds()),
Attributes: map[string]interface{}{
DisplayUnitsLong: Operations,
DisplayUnitsShort: OperationsShort,
DisplayMin: "0",
},
},
Measurement{
Name: fmt.Sprintf("%s.%s", name, "15min"),
Value: ms.Rate15(),
Period: int64(rep.Interval.Seconds()),
Attributes: map[string]interface{}{
DisplayUnitsLong: Operations,
DisplayUnitsShort: OperationsShort,
DisplayMin: "0",
},
},
)
case metrics.Timer:
ms := m.Snapshot()
measurement[Name] = name
measurement[Value] = float64(ms.Count())
snapshot.Counters = append(snapshot.Counters, measurement)
if ms.Count() > 0 {
libratoName := fmt.Sprintf("%s.%s", name, "timer.mean")
gauges := make([]Measurement, histogramGaugeCount)
gauges[0] = Measurement{
Name: libratoName,
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: ms.Percentile(p),
Period: int64(rep.Interval.Seconds()),
Attributes: rep.TimerAttributes,
}
}
snapshot.Gauges = append(snapshot.Gauges, gauges...)
snapshot.Gauges = append(snapshot.Gauges,
Measurement{
Name: fmt.Sprintf("%s.%s", name, "rate.1min"),
Value: ms.Rate1(),
Period: int64(rep.Interval.Seconds()),
Attributes: map[string]interface{}{
DisplayUnitsLong: Operations,
DisplayUnitsShort: OperationsShort,
DisplayMin: "0",
},
},
Measurement{
Name: fmt.Sprintf("%s.%s", name, "rate.5min"),
Value: ms.Rate5(),
Period: int64(rep.Interval.Seconds()),
Attributes: map[string]interface{}{
DisplayUnitsLong: Operations,
DisplayUnitsShort: OperationsShort,
DisplayMin: "0",
},
},
Measurement{
Name: fmt.Sprintf("%s.%s", name, "rate.15min"),
Value: ms.Rate15(),
Period: int64(rep.Interval.Seconds()),
Attributes: map[string]interface{}{
DisplayUnitsLong: Operations,
DisplayUnitsShort: OperationsShort,
DisplayMin: "0",
},
},
)
}
}
})
return
}