backend-observability
Skill2 starsChanged 5 months agoMITView on GitHub
Logs estructurados, métricas de los Four Golden Signals, trazas distribuidas, correlación por request-id, niveles correctos, qué NUNCA loguear. Cargá al instrumentar un servicio o investigar un incidente.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- kensaurus/cursor-kenjitheir own version, first seen Oct 1, 2026
- kensaurus/skillstheir own version, first seen Oct 1, 2026
- omonuj/claude-skillstheir own version, first seen Jul 23, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- kensaurus/cursor-kenji★ 9same name
- kensaurus/skills★ 9same name
- omonuj/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository25 other files it gives its agents.
- accessibility-wcagSkill
- agent-orchestrationSkill
- backend-api-designSkill
- backend-data-modelingSkill
- backend-performanceSkill
- blossom-carouselSkill
- browser-qaSkill
- commit-hygieneSkill
- discovery-intakeSkill
- frontend-component-patternsSkill
- geo-seoSkill
- gh-cliSkill
- glosario-vivoSkill
- llm-councilSkill
- pedido-claroSkill
- ponytailSkill
- reporte-comprensibleSkill
- sdd-flowSkill
- secure-by-defaultSkill
- security-authn-authzSkill
- security-owaspSkill
- security-secretsSkill
- security-supply-chainSkill
- turtle-protocolSkill
- ui-ux-pro-maxSkill
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 39
Un log por evento de negocio: No logs decorativos cada 3 líneas.
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Backend observability
Cuándo usar
- Crear un servicio nuevo o un endpoint sensible.
- Investigar un incidente y faltan datos.
- Definir alertas o dashboards.
Pilares
- Logs — eventos discretos, con contexto, estructurados (JSON).
- Métricas — series temporales agregables.
- Trazas — el camino de un request por varios servicios.
Sin los tres, falta una pata: logs sin métricas no escalan, métricas sin trazas no explican causa.
Logs
- Estructurados (
{"ts":..., "level":..., "msg":..., "request_id":..., ...}), nunca strings sueltos. - Niveles bien usados:
DEBUG— detalle de desarrollo. Off en producción.INFO— eventos esperados (request servido, job ok). Bajo volumen.WARN— anomalía recuperada (reintento, fallback). Mirar tendencias.ERROR— fallo no recuperado en un request/job. Debería alertar si sube.FATAL— el proceso no puede continuar.
- Un log por evento de negocio. No logs decorativos cada 3 líneas.
request_idpropagado en todo log de la cadena. Si no podés correlacionar, no podés debuggear.
Qué NUNCA loguear
- Contraseñas, tokens, claves, header
Authorization, cookies de sesión. - PII completa: número de tarjeta, RUT/DNI, mail, teléfono. Hashear o truncar si hace falta.
- Cuerpo de request/response salvo en debug local explícito.
- Stack traces a clientes (sí internamente).
Métricas — Four Golden Signals
Para todo servicio:
- Latency — p50, p95, p99 por endpoint. Promedio miente.
- Traffic — requests/seg, por ruta y método.
- Errors — tasa de 5xx (y 4xx interesantes como 429).
- Saturation — CPU, memoria, conexiones a BD, cola.
Para jobs: éxito/falla/duración/cola.
Convenciones
- Histogramas, no solo contadores, para latencias.
- Cardinalidad bajo control:
user_idcomo label rompe Prometheus. - Nombres consistentes:
http_request_duration_seconds,db_query_duration_seconds.
Trazas
- OpenTelemetry como estándar. Cliente + auto-instrumentación + exportador a Jaeger/Tempo/Honeycomb.
- Propagá
traceparententre servicios. - Span por unidad de trabajo significativa (request, query, llamada externa, job).
- Atributos útiles:
http.route,db.statement(sanitizado),messaging.destination.
Health checks
/health/live— el proceso responde. Para Kubernetes liveness./health/ready— listo para tráfico (BD ok, dependencias críticas ok). Para readiness.- No mezclar.
liveno debería tocar BD.
Alertas
- Alertá sobre síntomas del usuario (latencia alta, error rate >X%), no sobre causas (CPU 80%).
- Toda alerta tiene runbook con: qué significa, qué chequear, qué hacer.
- Umbrales basados en SLO, no en intuición.
Reglas duras
- Sin
console.logniprintdirectos en producción. - Sin loguear en loop apretado sin sampling.
- Sin métrica nueva sin dashboard. Si no se mira, no existe.
- Sin alerta sin runbook. Ruido garantizado.
Validación
- Generá un request en local y seguilo por log, métrica y traza.
- Caos test: matá la BD, ¿la alerta dispara en <5 min?
- Revisión periódica: ¿qué alertas dispararon? ¿cuáles fueron ruido?
Relacionadas
[[backend-api-design]] · [[backend-performance]] · [[security-secrets]]
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

