telemetry/logging: share the slog setup every service builds by hand
I maintainer di solito rispondono entro 1 giorno
Nessuno ha ancora preso questa issue.
Valutazione
- Difficoltà
- 5/5
- Tempo stimato
- Più di una settimana
- Idoneità per principianti
- 35/100
- Tipo di issue
- Funzionalità
- Chiarezza
- Abbastanza chiara
- Stato di attività
- Attiva
- Stack tecnologico
- go
- Ambito
- observability
Direzione di ricerca
Start with the existing telemetry/logging package and the dependent work in #359, then compare the logging implementations named in the issue: openndx-core internal/logging, nsw-srilanka internal/bootstrap/logging.go, and agency backend/internal/logging. Review openndx-core#708 for its logging rules and tests. Done means a shared setup covers the listed configuration, trace IDs, optional OTLP export with level filtering, common service attributes, and the default logger and correlation ID wiring; several design decisions and consumer migrations remain.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
Current Limitation
Every OpenNSW service builds its slog setup by hand. All of them build the same chain: base handler → trace ID decorator → (optional) OTLP export → slog.SetDefault → httputil.CorrelationIDFunc. Each one does it a little differently:
openndx-core internal/logging |
nsw-srilanka internal/bootstrap/logging.go |
agency backend/internal/logging |
|
|---|---|---|---|
| Level from config | logging.level |
server.logLevel |
Not configurable |
| Format | json or text, configurable |
text only |
json only |
service / environment / version on every line |
Yes | No | No |
Trace ID decorator (trace/logging) |
No | Yes | Yes |
httputil.CorrelationIDFunc wired |
No | Yes | Yes |
OTLP log export (otelslog + slog.NewMultiHandler) |
No | Yes | No |
Level filter in front of otelslog |
— | Yes | — |
AddSource at debug |
No | Yes | No |
So:
- A combined log stream can't be filtered by service, because only one of the three adds
service, and its value isn't tied to the OTelservice.name. - Fixes stay where they were found. nsw-srilanka found that
otelslogreports every level as enabled, so behindslog.NewMultiHandlerthe multi-handler'sEnabledis true for DEBUG even when the configured level is INFO. DEBUG records were exported and debug-gated payload logging turned on. ItslevelFilterfixes this. Any other service that adds OTLP log export will hit the same bug. - Config keys and defaults differ (
logging.levelvsserver.logLevel,jsonvstext), so operators configure each service differently.
Suggested Improvement
Add the logging setup to telemetry/logging, beside the trace_id / span_id decorator from #359.
package logging // github.com/OpenNSW/core/telemetry/logging
// Config is the logging section of a service's config file.
type Config struct {
// Level is debug, info, warn or error. Default info.
Level string `yaml:"level"`
// Format is json or text. Default json.
Format string `yaml:"format"`
}
// DefaultConfig returns the values used for keys the config file leaves out.
func DefaultConfig() Config
// Validate reports an unknown level or format.
func (c Config) Validate() error
// Setup builds the logger, sets it as slog's default and returns it. cfg must be valid.
func Setup(cfg Config, p *telemetry.Provider) *slog.Logger
// NewHandler adds trace_id and span_id from the span in ctx (#359).
func NewHandler(next slog.Handler) slog.Handler
Setup builds:
slog.JSONHandlerorslog.TextHandleron stdout, atcfg.Level, withAddSourceat debugNewHandleraround it, fortrace_id/span_id- When
pexports logs over OTLP:p.SlogHandler()behind a level filter andNewHandler, fanned out withslog.NewMultiHandler service,environmentandversionattributes, taken from the same values as theProvider's resource (service.name,deployment.environment,service.version), so the log field and the OTel attribute can't driftslog.SetDefault, andhttputil.CorrelationIDFunc = telemetry.TraceID
Each main then has:
p, err := telemetry.Start(ctx, cfg.Telemetry)
// ...
logging.Setup(cfg.Logging, p)
Service-specific behavior stays in the service, such as nsw-srilanka's warning that audit records are suppressed above info.
Consumers to migrate
- openndx/openndx-core:
internal/logging(openndx/openndx-core#708) is deleted, andcmd/<service>/main.gocallslogging.Setup. - OpenNSW/nsw-srilanka:
internal/bootstrap/logging.go, includinglevelFilter.server.logLeveleither moves to aloggingsection or is mapped intologging.Configby its config code. - OpenNSW/agency:
backend/internal/logging.
Decisions to settle in review
- Default format:
json(openndx-core, agency) ortext(nsw-srilanka). Suggestjsonfor production, since log pipelines parse it. - Config key:
logging.level/logging.formatas one section, or leave each service to map its own keys intoConfig. - Is
Setupusable without telemetry? For example withp == nilit writes stdout only and adds no trace IDs. Or does every service starttelemetryfirst?
Version
github.com/OpenNSW/core/trace v0.2.0
Additional Context
- Depends on #359 for
telemetry.Provider,telemetry.TraceIDandtelemetry/logging.NewHandler. - openndx/openndx-core#708 has a description of openndx's logging rules (one setup, one global logger,
...Contextcalls, common fields). Itsinternal/loggingpackage is the starting point forConfig,Validateand the common fields.
- Lingua principale
- Go
- Stelle
- 6
- Fork
- 8
- Merge medio
- 21h 36m
- PR unite (30g)
- 42
Preparare l'ambiente
- Nessun Dockerfile né file Docker Compose
- Ha un modello di pull request
- Leggi la guida per i contributori
Come iniziare
- Leggi tutta la issue e poi la guida ai contributi del progetto.
- Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
- Fai un fork del repository e lavora su un branch.
- Apri una pull request che faccia riferimento al numero della issue.
Altre issue di OpenNSW/core
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 74/100
I maintainer di solito rispondono entro 1 giorno
-
payment: test fixtures hardcode real gateway names (govpay/lankapay) instead of generic placeholdersAperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 88/100
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
I maintainer di solito rispondono entro 1 giorno
-
nsw-sl cleanup Type/Improvement
Difficoltà 3/5 1-2 giorni Idoneità per principianti 48/100
I maintainer di solito rispondono entro 1 giorno
-
remote: let the host inject a logger, and log with one consistent policyForse già presa @mushrafmim l’ha presa 5 giorni fa. ApertaType/Improvement
Difficoltà 4/5 3-5 giorni Idoneità per principianti 22/100
I maintainer di solito rispondono entro 1 giorno
Tutte le issue di OpenNSW/core
Issue simili
-
agent-research-finding chore
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
jordansmall/spindrift#7370 ·
I maintainer di solito rispondono entro 1 giorno
-
☢️ Bug v3
Difficoltà 2/5 1-3 ore Idoneità per principianti 82/100
I maintainer di solito rispondono entro 1 giorno
-
The repo's gates check is clean and runs need no swapsForse già presa @turbot-agent-fleet l’ha presa oggi. Apertaagent:ready
Difficoltà 2/5 1-3 ore Idoneità per principianti 64/100
turbot/tailpipe#677 · 2 commenti ·
I maintainer di solito rispondono entro 1 giorno
-
needs-kind needs-triage
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
I maintainer di solito rispondono entro 2 giorni
-
enhancement
Difficoltà 2/5 1-3 ore Idoneità per principianti 70/100
crossplane-contrib/provider-keycloak#806 ·
I maintainer di solito rispondono entro 1 giorno