Files
syncthing/cmd/infra/strelaypoolsrv/stats.go
T
Jakob BorgandRoss Smith II 836045ee87 feat: switch logging framework (#10220)
This updates our logging framework from legacy freetext strings using
the `log` package to structured log entries using `log/slog`. I have
updated all INFO or higher level entries, but not yet DEBUG (😓)... So,
at a high level:

There is a slight change in log levels, effectively adding a new warning
level:

- DEBUG is still debug (ideally not for users but developers, though
this is something we need to work on)
- INFO is still info, though I've added more data here, effectively
making Syncthing more verbose by default (more on this below)
- WARNING is a new log level that is different from the _old_ WARNING
(more below)
- ERROR is what was WARNING before -- problems that must be dealt with,
and also bubbled as a popup in the GUI.

A new feature is that the logging level can be set per package to
something other than just debug or info, and hence I feel that we can
add a bit more things into INFO while moving some (in fact, most)
current INFO level warnings into WARNING. For example, I think it's
justified to get a log of synced files in INFO and sync failures in
WARNING. These are things that have historically been tricky to debug
properly, and having more information by default will be useful to many,
while still making it possible get close to told level of inscrutability
by setting the log level to WARNING. I'd like to get to a stage where
DEBUG is never necessary to just figure out what's going on, as opposed
to trying to narrow down a likely bug.

Code wise:

- Our logging object, generally known as `l` in each package, is now a
new adapter object that provides the old API on top of the newer one.
(This should go away once all old log entries are migrated.) This is
only for `l.Debugln` and `l.Debugf`.
- There is a new level tracker that keeps the log level for each
package.
- There is a nested setup of handlers, since the structure mandated by
`log/slog` is slightly convoluted (imho). We do this because we need to
do formatting at a "medium" level internally so we can buffer log lines
in text format but with separate timestamp and log level for the API/GUI
to consume.
- The `debug` API call becomes a `loglevels` API call, which can set the
log level to `DEBUG`, `INFO`, `WARNING` or `ERROR` per package. The GUI
is updated to handle this.
- Our custom `sync` package provided some debugging of mutexes quite
strongly integrated into the old logging framework, only turned on when
`STTRACE` was set to certain values at startup, etc. It's been a long
time since this has been useful; I removed it.
- The `STTRACE` env var remains and can be used the same way as before,
while additionally permitting specific log levels to be specified,
`STTRACE=model:WARN,scanner:DEBUG`.
- There is a new command line option `--log-level=INFO` to set the
default log level.
- The command line options `--log-flags` and `--verbose` go away, but
are currently retained as hidden & ignored options since we set them by
default in some of our startup examples and Syncthing would otherwise
fail to start.

Sample format messages:

```
2009-02-13 23:31:30 INF A basic info line (attr1="val with spaces" attr2=2 attr3="val\"quote" a=a log.pkg=slogutil)
2009-02-13 23:31:30 INF An info line with grouped values (attr1=val1 foo.attr2=2 foo.bar.attr3=3 a=a log.pkg=slogutil)
2009-02-13 23:31:30 INF An info line with grouped values via logger (foo.attr1=val1 foo.attr2=2 a=a log.pkg=slogutil)
2009-02-13 23:31:30 INF An info line with nested grouped values via logger (bar.foo.attr1=val1 bar.foo.attr2=2 a=a log.pkg=slogutil)
2009-02-13 23:31:30 WRN A warning entry (a=a log.pkg=slogutil)
2009-02-13 23:31:30 ERR An error (a=a log.pkg=slogutil)
```

---------

Co-authored-by: Ross Smith II <ross@smithii.com>
2025-08-07 11:19:36 +02:00

246 lines
8.1 KiB
Go

// Copyright (C) 2018 Audrius Butkevicius and Contributors (see the CONTRIBUTORS file).
package main
import (
"encoding/json"
"net"
"net/http"
"sync"
"time"
"github.com/prometheus/client_golang/prometheus"
)
var (
statusClient = http.Client{
Timeout: 5 * time.Second,
}
apiRequestsTotal = makeCounter("api_requests_total", "Number of API requests.", "type", "result")
apiRequestsSeconds = makeSummary("api_requests_seconds", "Latency of API requests.", "type")
relayTestsTotal = makeCounter("tests_total", "Number of relay tests.", "result")
relayTestActionsSeconds = makeSummary("test_actions_seconds", "Latency of relay test actions.", "type")
locationLookupSeconds = makeSummary("location_lookup_seconds", "Latency of location lookups.").WithLabelValues()
metricsRequestsSeconds = makeSummary("metrics_requests_seconds", "Latency of metric requests.").WithLabelValues()
scrapeSeconds = makeSummary("relay_scrape_seconds", "Latency of metric scrapes from remote relays.", "result")
relayUptime = makeGauge("relay_uptime", "Uptime of relay", "relay")
relayPendingSessionKeys = makeGauge("relay_pending_session_keys", "Number of pending session keys (two keys per session, one per each side of the connection)", "relay")
relayActiveSessions = makeGauge("relay_active_sessions", "Number of sessions that are happening, a session contains two parties", "relay")
relayConnections = makeGauge("relay_connections", "Number of devices connected to the relay", "relay")
relayProxies = makeGauge("relay_proxies", "Number of active proxy routines sending data between peers (two proxies per session, one for each way)", "relay")
relayBytesProxied = makeGauge("relay_bytes_proxied", "Number of bytes proxied by the relay", "relay")
relayGoRoutines = makeGauge("relay_go_routines", "Number of Go routines in the process", "relay")
relaySessionRate = makeGauge("relay_session_rate", "Rate applied per session", "relay")
relayGlobalRate = makeGauge("relay_global_rate", "Global rate applied on the whole relay", "relay")
relayBuildInfo = makeGauge("relay_build_info", "Build information about a relay", "relay", "go_version", "go_os", "go_arch")
relayLocationInfo = makeGauge("relay_location_info", "Location information about a relay", "relay", "city", "country", "continent")
lastStats = make(map[string]stats)
)
func makeGauge(name string, help string, labels ...string) *prometheus.GaugeVec {
gauge := prometheus.NewGaugeVec(
prometheus.GaugeOpts{
Namespace: "syncthing",
Subsystem: "relaypoolsrv",
Name: name,
Help: help,
},
labels,
)
prometheus.MustRegister(gauge)
return gauge
}
func makeSummary(name string, help string, labels ...string) *prometheus.SummaryVec {
summary := prometheus.NewSummaryVec(
prometheus.SummaryOpts{
Namespace: "syncthing",
Subsystem: "relaypoolsrv",
Name: name,
Help: help,
Objectives: map[float64]float64{0.5: 0.05, 0.9: 0.01, 0.99: 0.001},
},
labels,
)
prometheus.MustRegister(summary)
return summary
}
func makeCounter(name string, help string, labels ...string) *prometheus.CounterVec {
counter := prometheus.NewCounterVec(
prometheus.CounterOpts{
Namespace: "syncthing",
Subsystem: "relaypoolsrv",
Name: name,
Help: help,
},
labels,
)
prometheus.MustRegister(counter)
return counter
}
func statsRefresher(interval time.Duration) {
ticker := time.NewTicker(interval)
for range ticker.C {
refreshStats()
}
}
type statsFetchResult struct {
relay *relay
stats *stats
}
func refreshStats() {
mut.RLock()
relays := append(permanentRelays, knownRelays...)
mut.RUnlock()
now := time.Now()
var wg sync.WaitGroup
results := make(chan statsFetchResult, len(relays))
for _, rel := range relays {
wg.Add(1)
go func(rel *relay) {
t0 := time.Now()
stats := fetchStats(rel)
duration := time.Since(t0).Seconds()
result := "success"
if stats == nil {
result = "failed"
}
scrapeSeconds.WithLabelValues(result).Observe(duration)
results <- statsFetchResult{
relay: rel,
stats: fetchStats(rel),
}
wg.Done()
}(rel)
}
wg.Wait()
close(results)
mut.Lock()
relayBuildInfo.Reset()
relayLocationInfo.Reset()
for result := range results {
result.relay.StatsRetrieved = now
result.relay.Stats = result.stats
if result.stats == nil {
deleteMetrics(result.relay.uri.Host)
} else {
updateMetrics(result.relay.uri.Host, *result.stats, result.relay.Location)
}
}
mut.Unlock()
}
func fetchStats(relay *relay) *stats {
statusAddr := relay.uri.Query().Get("statusAddr")
if statusAddr == "" {
statusAddr = ":22070"
}
statusHost, statusPort, err := net.SplitHostPort(statusAddr)
if err != nil {
return nil
}
if statusHost == "" {
if host, _, err := net.SplitHostPort(relay.uri.Host); err != nil {
return nil
} else {
statusHost = host
}
}
url := "http://" + net.JoinHostPort(statusHost, statusPort) + "/status"
response, err := statusClient.Get(url)
if err != nil {
return nil
}
var stats stats
if err := json.NewDecoder(response.Body).Decode(&stats); err != nil {
return nil
}
return &stats
}
func updateMetrics(host string, stats stats, location location) {
if stats.GoVersion != "" || stats.GoOS != "" || stats.GoArch != "" {
relayBuildInfo.WithLabelValues(host, stats.GoVersion, stats.GoOS, stats.GoArch).Add(1)
}
if location.City != "" || location.Country != "" || location.Continent != "" {
relayLocationInfo.WithLabelValues(host, location.City, location.Country, location.Continent).Add(1)
}
if lastStat, ok := lastStats[host]; ok {
stats = mergeStats(stats, lastStat)
}
relayUptime.WithLabelValues(host).Set(float64(stats.UptimeSeconds))
relayPendingSessionKeys.WithLabelValues(host).Set(float64(stats.PendingSessionKeys))
relayActiveSessions.WithLabelValues(host).Set(float64(stats.ActiveSessions))
relayConnections.WithLabelValues(host).Set(float64(stats.Connections))
relayProxies.WithLabelValues(host).Set(float64(stats.Proxies))
relayBytesProxied.WithLabelValues(host).Set(float64(stats.BytesProxied))
relayGoRoutines.WithLabelValues(host).Set(float64(stats.GoRoutines))
relaySessionRate.WithLabelValues(host).Set(float64(stats.Options.SessionRate))
relayGlobalRate.WithLabelValues(host).Set(float64(stats.Options.GlobalRate))
lastStats[host] = stats
}
func deleteMetrics(host string) {
relayUptime.DeleteLabelValues(host)
relayPendingSessionKeys.DeleteLabelValues(host)
relayActiveSessions.DeleteLabelValues(host)
relayConnections.DeleteLabelValues(host)
relayProxies.DeleteLabelValues(host)
relayBytesProxied.DeleteLabelValues(host)
relayGoRoutines.DeleteLabelValues(host)
relaySessionRate.DeleteLabelValues(host)
relayGlobalRate.DeleteLabelValues(host)
delete(lastStats, host)
}
// Due to some unexplainable behaviour, some of the numbers sometimes travel slightly backwards (by less than 1%)
// This happens between scrapes, which is 30s, so this can't be a race.
// This causes prometheus to assume a "rate reset", hence causes phenomenal spikes.
// One of the number that moves backwards is BytesProxied, which atomically increments a counter with numeric value
// returned by net.Conn.Read(). I don't think that can return a negative value, so I have no idea what's going on.
func mergeStats(new stats, old stats) stats {
new.UptimeSeconds = mergeValue(new.UptimeSeconds, old.UptimeSeconds)
new.PendingSessionKeys = mergeValue(new.PendingSessionKeys, old.PendingSessionKeys)
new.ActiveSessions = mergeValue(new.ActiveSessions, old.ActiveSessions)
new.Connections = mergeValue(new.Connections, old.Connections)
new.Proxies = mergeValue(new.Proxies, old.Proxies)
new.BytesProxied = mergeValue(new.BytesProxied, old.BytesProxied)
new.GoRoutines = mergeValue(new.GoRoutines, old.GoRoutines)
new.Options.SessionRate = mergeValue(new.Options.SessionRate, old.Options.SessionRate)
new.Options.GlobalRate = mergeValue(new.Options.GlobalRate, old.Options.GlobalRate)
return new
}
func mergeValue(new, old int) int {
if new >= old {
return new // normal increase
}
if float64(new) > 0.99*float64(old) {
return old // slight backward movement
}
return new // reset (relay restart)
}