2018-02-23 03:56:08 -06:00
|
|
|
// Go port of Coda Hale's Metrics library
|
2015-07-06 19:54:22 -05:00
|
|
|
//
|
2018-02-23 03:56:08 -06:00
|
|
|
// <https://github.com/rcrowley/go-metrics>
|
2015-07-06 19:54:22 -05:00
|
|
|
//
|
2018-02-23 03:56:08 -06:00
|
|
|
// Coda Hale's original work: <https://github.com/codahale/metrics>
|
2025-01-14 07:16:15 -06:00
|
|
|
|
2015-06-27 10:12:58 -05:00
|
|
|
package metrics
|
|
|
|
|
|
|
|
import (
|
2022-11-11 06:16:13 -06:00
|
|
|
"runtime/metrics"
|
|
|
|
"runtime/pprof"
|
2015-06-27 10:12:58 -05:00
|
|
|
"time"
|
|
|
|
)
|
|
|
|
|
metrics, cmd/geth: change init-process of metrics (#30814)
This PR modifies how the metrics library handles `Enabled`: previously,
the package `init` decided whether to serve real metrics or just
dummy-types.
This has several drawbacks:
- During pkg init, we need to determine whether metrics are enabled or
not. So we first hacked in a check if certain geth-specific
commandline-flags were enabled. Then we added a similar check for
geth-env-vars. Then we almost added a very elaborate check for
toml-config-file, plus toml parsing.
- Using "real" types and dummy types interchangeably means that
everything is hidden behind interfaces. This has a performance penalty,
and also it just adds a lot of code.
This PR removes the interface stuff, uses concrete types, and allows for
the setting of Enabled to happen later. It is still assumed that
`metrics.Enable()` is invoked early on.
The somewhat 'heavy' operations, such as ticking meters and exp-decay,
now checks the enable-flag to prevent resource leak.
The change may be large, but it's mostly pretty trivial, and from the
last time I gutted the metrics, I ensured that we have fairly good test
coverage.
---------
Co-authored-by: Felix Lange <fjl@twurst.com>
2024-12-10 06:27:29 -06:00
|
|
|
var (
|
|
|
|
metricsEnabled = false
|
|
|
|
)
|
2023-09-14 05:56:06 -05:00
|
|
|
|
metrics, cmd/geth: change init-process of metrics (#30814)
This PR modifies how the metrics library handles `Enabled`: previously,
the package `init` decided whether to serve real metrics or just
dummy-types.
This has several drawbacks:
- During pkg init, we need to determine whether metrics are enabled or
not. So we first hacked in a check if certain geth-specific
commandline-flags were enabled. Then we added a similar check for
geth-env-vars. Then we almost added a very elaborate check for
toml-config-file, plus toml parsing.
- Using "real" types and dummy types interchangeably means that
everything is hidden behind interfaces. This has a performance penalty,
and also it just adds a lot of code.
This PR removes the interface stuff, uses concrete types, and allows for
the setting of Enabled to happen later. It is still assumed that
`metrics.Enable()` is invoked early on.
The somewhat 'heavy' operations, such as ticking meters and exp-decay,
now checks the enable-flag to prevent resource leak.
The change may be large, but it's mostly pretty trivial, and from the
last time I gutted the metrics, I ensured that we have fairly good test
coverage.
---------
Co-authored-by: Felix Lange <fjl@twurst.com>
2024-12-10 06:27:29 -06:00
|
|
|
// Enabled is checked by functions that are deemed 'expensive', e.g. if a
|
|
|
|
// meter-type does locking and/or non-trivial math operations during update.
|
|
|
|
func Enabled() bool {
|
|
|
|
return metricsEnabled
|
|
|
|
}
|
2019-03-25 03:01:18 -05:00
|
|
|
|
metrics, cmd/geth: change init-process of metrics (#30814)
This PR modifies how the metrics library handles `Enabled`: previously,
the package `init` decided whether to serve real metrics or just
dummy-types.
This has several drawbacks:
- During pkg init, we need to determine whether metrics are enabled or
not. So we first hacked in a check if certain geth-specific
commandline-flags were enabled. Then we added a similar check for
geth-env-vars. Then we almost added a very elaborate check for
toml-config-file, plus toml parsing.
- Using "real" types and dummy types interchangeably means that
everything is hidden behind interfaces. This has a performance penalty,
and also it just adds a lot of code.
This PR removes the interface stuff, uses concrete types, and allows for
the setting of Enabled to happen later. It is still assumed that
`metrics.Enable()` is invoked early on.
The somewhat 'heavy' operations, such as ticking meters and exp-decay,
now checks the enable-flag to prevent resource leak.
The change may be large, but it's mostly pretty trivial, and from the
last time I gutted the metrics, I ensured that we have fairly good test
coverage.
---------
Co-authored-by: Felix Lange <fjl@twurst.com>
2024-12-10 06:27:29 -06:00
|
|
|
// Enable enables the metrics system.
|
|
|
|
// The Enabled-flag is expected to be set, once, during startup, but toggling off and on
|
|
|
|
// is not supported.
|
|
|
|
//
|
|
|
|
// Enable is not safe to call concurrently. You need to call this as early as possible in
|
|
|
|
// the program, before any metrics collection will happen.
|
|
|
|
func Enable() {
|
|
|
|
metricsEnabled = true
|
2015-06-29 08:11:01 -05:00
|
|
|
}
|
|
|
|
|
2022-11-11 06:16:13 -06:00
|
|
|
var threadCreateProfile = pprof.Lookup("threadcreate")
|
|
|
|
|
|
|
|
type runtimeStats struct {
|
|
|
|
GCPauses *metrics.Float64Histogram
|
|
|
|
GCAllocBytes uint64
|
|
|
|
GCFreedBytes uint64
|
|
|
|
|
|
|
|
MemTotal uint64
|
|
|
|
HeapObjects uint64
|
|
|
|
HeapFree uint64
|
|
|
|
HeapReleased uint64
|
|
|
|
HeapUnused uint64
|
|
|
|
|
|
|
|
Goroutines uint64
|
|
|
|
SchedLatency *metrics.Float64Histogram
|
|
|
|
}
|
|
|
|
|
|
|
|
var runtimeSamples = []metrics.Sample{
|
|
|
|
{Name: "/gc/pauses:seconds"}, // histogram
|
|
|
|
{Name: "/gc/heap/allocs:bytes"},
|
|
|
|
{Name: "/gc/heap/frees:bytes"},
|
|
|
|
{Name: "/memory/classes/total:bytes"},
|
|
|
|
{Name: "/memory/classes/heap/objects:bytes"},
|
|
|
|
{Name: "/memory/classes/heap/free:bytes"},
|
|
|
|
{Name: "/memory/classes/heap/released:bytes"},
|
|
|
|
{Name: "/memory/classes/heap/unused:bytes"},
|
|
|
|
{Name: "/sched/goroutines:goroutines"},
|
|
|
|
{Name: "/sched/latencies:seconds"}, // histogram
|
|
|
|
}
|
|
|
|
|
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 12:13:47 -05:00
|
|
|
func ReadRuntimeStats() *runtimeStats {
|
|
|
|
r := new(runtimeStats)
|
|
|
|
readRuntimeStats(r)
|
|
|
|
return r
|
|
|
|
}
|
|
|
|
|
2022-11-11 06:16:13 -06:00
|
|
|
func readRuntimeStats(v *runtimeStats) {
|
|
|
|
metrics.Read(runtimeSamples)
|
|
|
|
for _, s := range runtimeSamples {
|
|
|
|
// Skip invalid/unknown metrics. This is needed because some metrics
|
|
|
|
// are unavailable in older Go versions, and attempting to read a 'bad'
|
|
|
|
// metric panics.
|
|
|
|
if s.Value.Kind() == metrics.KindBad {
|
|
|
|
continue
|
|
|
|
}
|
|
|
|
|
|
|
|
switch s.Name {
|
|
|
|
case "/gc/pauses:seconds":
|
|
|
|
v.GCPauses = s.Value.Float64Histogram()
|
|
|
|
case "/gc/heap/allocs:bytes":
|
|
|
|
v.GCAllocBytes = s.Value.Uint64()
|
|
|
|
case "/gc/heap/frees:bytes":
|
|
|
|
v.GCFreedBytes = s.Value.Uint64()
|
|
|
|
case "/memory/classes/total:bytes":
|
|
|
|
v.MemTotal = s.Value.Uint64()
|
|
|
|
case "/memory/classes/heap/objects:bytes":
|
|
|
|
v.HeapObjects = s.Value.Uint64()
|
|
|
|
case "/memory/classes/heap/free:bytes":
|
|
|
|
v.HeapFree = s.Value.Uint64()
|
|
|
|
case "/memory/classes/heap/released:bytes":
|
|
|
|
v.HeapReleased = s.Value.Uint64()
|
|
|
|
case "/memory/classes/heap/unused:bytes":
|
|
|
|
v.HeapUnused = s.Value.Uint64()
|
|
|
|
case "/sched/goroutines:goroutines":
|
|
|
|
v.Goroutines = s.Value.Uint64()
|
|
|
|
case "/sched/latencies:seconds":
|
|
|
|
v.SchedLatency = s.Value.Float64Histogram()
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// CollectProcessMetrics periodically collects various metrics about the running process.
|
2015-06-27 10:12:58 -05:00
|
|
|
func CollectProcessMetrics(refresh time.Duration) {
|
2015-06-29 08:11:01 -05:00
|
|
|
// Short circuit if the metrics system is disabled
|
metrics, cmd/geth: change init-process of metrics (#30814)
This PR modifies how the metrics library handles `Enabled`: previously,
the package `init` decided whether to serve real metrics or just
dummy-types.
This has several drawbacks:
- During pkg init, we need to determine whether metrics are enabled or
not. So we first hacked in a check if certain geth-specific
commandline-flags were enabled. Then we added a similar check for
geth-env-vars. Then we almost added a very elaborate check for
toml-config-file, plus toml parsing.
- Using "real" types and dummy types interchangeably means that
everything is hidden behind interfaces. This has a performance penalty,
and also it just adds a lot of code.
This PR removes the interface stuff, uses concrete types, and allows for
the setting of Enabled to happen later. It is still assumed that
`metrics.Enable()` is invoked early on.
The somewhat 'heavy' operations, such as ticking meters and exp-decay,
now checks the enable-flag to prevent resource leak.
The change may be large, but it's mostly pretty trivial, and from the
last time I gutted the metrics, I ensured that we have fairly good test
coverage.
---------
Co-authored-by: Felix Lange <fjl@twurst.com>
2024-12-10 06:27:29 -06:00
|
|
|
if !metricsEnabled {
|
2015-06-29 08:11:01 -05:00
|
|
|
return
|
|
|
|
}
|
2022-11-11 06:16:13 -06:00
|
|
|
|
2015-06-27 10:12:58 -05:00
|
|
|
// Create the various data collectors
|
2019-06-17 02:53:17 -05:00
|
|
|
var (
|
2022-11-11 06:16:13 -06:00
|
|
|
cpustats = make([]CPUStats, 2)
|
|
|
|
diskstats = make([]DiskStats, 2)
|
|
|
|
rstats = make([]runtimeStats, 2)
|
|
|
|
)
|
|
|
|
|
|
|
|
// This scale factor is used for the runtime's time metrics. It's useful to convert to
|
|
|
|
// ns here because the runtime gives times in float seconds, but runtimeHistogram can
|
|
|
|
// only provide integers for the minimum and maximum values.
|
|
|
|
const secondsToNs = float64(time.Second)
|
2015-06-27 10:12:58 -05:00
|
|
|
|
2022-11-11 06:16:13 -06:00
|
|
|
// Define the various metrics to collect
|
|
|
|
var (
|
|
|
|
cpuSysLoad = GetOrRegisterGauge("system/cpu/sysload", DefaultRegistry)
|
|
|
|
cpuSysWait = GetOrRegisterGauge("system/cpu/syswait", DefaultRegistry)
|
|
|
|
cpuProcLoad = GetOrRegisterGauge("system/cpu/procload", DefaultRegistry)
|
2023-03-23 08:13:50 -05:00
|
|
|
cpuSysLoadTotal = GetOrRegisterCounterFloat64("system/cpu/sysload/total", DefaultRegistry)
|
|
|
|
cpuSysWaitTotal = GetOrRegisterCounterFloat64("system/cpu/syswait/total", DefaultRegistry)
|
|
|
|
cpuProcLoadTotal = GetOrRegisterCounterFloat64("system/cpu/procload/total", DefaultRegistry)
|
2022-11-11 06:16:13 -06:00
|
|
|
cpuThreads = GetOrRegisterGauge("system/cpu/threads", DefaultRegistry)
|
|
|
|
cpuGoroutines = GetOrRegisterGauge("system/cpu/goroutines", DefaultRegistry)
|
|
|
|
cpuSchedLatency = getOrRegisterRuntimeHistogram("system/cpu/schedlatency", secondsToNs, nil)
|
|
|
|
memPauses = getOrRegisterRuntimeHistogram("system/memory/pauses", secondsToNs, nil)
|
|
|
|
memAllocs = GetOrRegisterMeter("system/memory/allocs", DefaultRegistry)
|
|
|
|
memFrees = GetOrRegisterMeter("system/memory/frees", DefaultRegistry)
|
|
|
|
memTotal = GetOrRegisterGauge("system/memory/held", DefaultRegistry)
|
|
|
|
heapUsed = GetOrRegisterGauge("system/memory/used", DefaultRegistry)
|
|
|
|
heapObjects = GetOrRegisterGauge("system/memory/objects", DefaultRegistry)
|
2019-06-17 02:53:17 -05:00
|
|
|
diskReads = GetOrRegisterMeter("system/disk/readcount", DefaultRegistry)
|
|
|
|
diskReadBytes = GetOrRegisterMeter("system/disk/readdata", DefaultRegistry)
|
|
|
|
diskReadBytesCounter = GetOrRegisterCounter("system/disk/readbytes", DefaultRegistry)
|
|
|
|
diskWrites = GetOrRegisterMeter("system/disk/writecount", DefaultRegistry)
|
|
|
|
diskWriteBytes = GetOrRegisterMeter("system/disk/writedata", DefaultRegistry)
|
2018-07-09 07:11:49 -05:00
|
|
|
diskWriteBytesCounter = GetOrRegisterCounter("system/disk/writebytes", DefaultRegistry)
|
2019-06-17 02:53:17 -05:00
|
|
|
)
|
2022-11-11 06:16:13 -06:00
|
|
|
|
2023-03-06 17:29:48 -06:00
|
|
|
var lastCollectTime time.Time
|
|
|
|
|
2022-11-11 06:16:13 -06:00
|
|
|
// Iterate loading the different stats and updating the meters.
|
|
|
|
now, prev := 0, 1
|
|
|
|
for ; ; now, prev = prev, now {
|
2023-03-06 17:29:48 -06:00
|
|
|
// Gather CPU times.
|
2022-11-11 06:16:13 -06:00
|
|
|
ReadCPUStats(&cpustats[now])
|
2023-03-06 17:29:48 -06:00
|
|
|
collectTime := time.Now()
|
|
|
|
secondsSinceLastCollect := collectTime.Sub(lastCollectTime).Seconds()
|
|
|
|
lastCollectTime = collectTime
|
|
|
|
if secondsSinceLastCollect > 0 {
|
2023-03-23 08:13:50 -05:00
|
|
|
sysLoad := cpustats[now].GlobalTime - cpustats[prev].GlobalTime
|
|
|
|
sysWait := cpustats[now].GlobalWait - cpustats[prev].GlobalWait
|
|
|
|
procLoad := cpustats[now].LocalTime - cpustats[prev].LocalTime
|
2023-03-06 17:29:48 -06:00
|
|
|
// Convert to integer percentage.
|
2023-03-23 08:13:50 -05:00
|
|
|
cpuSysLoad.Update(int64(sysLoad / secondsSinceLastCollect * 100))
|
|
|
|
cpuSysWait.Update(int64(sysWait / secondsSinceLastCollect * 100))
|
|
|
|
cpuProcLoad.Update(int64(procLoad / secondsSinceLastCollect * 100))
|
|
|
|
// increment counters (ms)
|
|
|
|
cpuSysLoadTotal.Inc(sysLoad)
|
|
|
|
cpuSysWaitTotal.Inc(sysWait)
|
|
|
|
cpuProcLoadTotal.Inc(procLoad)
|
2023-03-06 17:29:48 -06:00
|
|
|
}
|
2022-11-11 06:16:13 -06:00
|
|
|
|
|
|
|
// Threads
|
2019-06-17 02:53:17 -05:00
|
|
|
cpuThreads.Update(int64(threadCreateProfile.Count()))
|
2022-11-11 06:16:13 -06:00
|
|
|
|
|
|
|
// Go runtime metrics
|
|
|
|
readRuntimeStats(&rstats[now])
|
|
|
|
|
|
|
|
cpuGoroutines.Update(int64(rstats[now].Goroutines))
|
|
|
|
cpuSchedLatency.update(rstats[now].SchedLatency)
|
|
|
|
memPauses.update(rstats[now].GCPauses)
|
|
|
|
|
|
|
|
memAllocs.Mark(int64(rstats[now].GCAllocBytes - rstats[prev].GCAllocBytes))
|
|
|
|
memFrees.Mark(int64(rstats[now].GCFreedBytes - rstats[prev].GCFreedBytes))
|
|
|
|
|
|
|
|
memTotal.Update(int64(rstats[now].MemTotal))
|
|
|
|
heapUsed.Update(int64(rstats[now].MemTotal - rstats[now].HeapUnused - rstats[now].HeapFree - rstats[now].HeapReleased))
|
|
|
|
heapObjects.Update(int64(rstats[now].HeapObjects))
|
|
|
|
|
|
|
|
// Disk
|
|
|
|
if ReadDiskStats(&diskstats[now]) == nil {
|
|
|
|
diskReads.Mark(diskstats[now].ReadCount - diskstats[prev].ReadCount)
|
|
|
|
diskReadBytes.Mark(diskstats[now].ReadBytes - diskstats[prev].ReadBytes)
|
|
|
|
diskWrites.Mark(diskstats[now].WriteCount - diskstats[prev].WriteCount)
|
|
|
|
diskWriteBytes.Mark(diskstats[now].WriteBytes - diskstats[prev].WriteBytes)
|
|
|
|
diskReadBytesCounter.Inc(diskstats[now].ReadBytes - diskstats[prev].ReadBytes)
|
|
|
|
diskWriteBytesCounter.Inc(diskstats[now].WriteBytes - diskstats[prev].WriteBytes)
|
2015-06-27 10:12:58 -05:00
|
|
|
}
|
2022-11-11 06:16:13 -06:00
|
|
|
|
2015-06-27 10:12:58 -05:00
|
|
|
time.Sleep(refresh)
|
|
|
|
}
|
|
|
|
}
|