agentsclimarketplace

Observability and monitoring

Skill agent-packs/registry/skills/observability-and-monitoring

Curated registry of Agent Packs — packs, skills, and plugins for AI coding agents

Install
npx -y skills add agent-packs/registry --skill observability-and-monitoring

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Instrument code with structured logs, metrics, and distributed traces. Use when adding observability to services, debugging production issues, or setting up monitoring.

The file declares its own license as Apache-2.0. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

2.6 KB, 530 tokens by cl100k_base, as published. Nobody here has run it

Observability and Monitoring

Instrument first, debug second. Good observability is built into code, not bolted on after incidents.

Structured Logging

  • Use structured (JSON) log output with consistent field names: level, timestamp, service, trace_id, span_id, message.
  • Log at the boundary of operations: request received, external call made, result returned.
  • Never log secrets, PII, or tokens — redact at the log site, not in a post-processor.
  • Use log levels correctly: DEBUG for trace/context, INFO for state transitions, WARN for degraded but recoverable, ERROR for failures needing attention.

Metrics

  • Instrument services with the RED method: Request rate, Error rate, Duration (latency p50/p95/p99).
  • For workers and queues use the USE method: Utilization, Saturation, Errors.
  • Name metrics with service_subsystem_unit_total or service_subsystem_unit_seconds conventions.
  • Add cardinality-bounded labels: status code, method, route template (not full URL), region. Never use user IDs or request IDs as labels.

Distributed Tracing

  • Propagate trace context (W3C TraceContext or B3) across all service boundaries: HTTP headers, message queue attributes, gRPC metadata.
  • Name spans with verb.noun conventions: db.query, cache.get, http.post.
  • Add span attributes at the point where the context is known.
  • Mark spans with error status and record the exception when errors propagate out.

Alerting and SLOs

  • Define SLOs before adding alerts. Alert on SLO burn rate, not raw thresholds.
  • Every alert must have a runbook link in its annotation.
  • Use multi-window burn rate alerts: short window for fast burn, long window for slow burn.

Health Endpoints

  • /healthz returns 200 only when the process is alive (liveness).
  • /readyz returns 200 only when all critical dependencies (database, cache) are reachable (readiness).
  • Log service startup and shutdown with version, commit SHA, and config hash.

Checklist

  • All external calls wrapped in a span with error and latency attributes.
  • Metrics include RED signals on every service boundary.
  • Trace context propagated across every async boundary.
  • /healthz and /readyz endpoints implemented and tested.
  • No secrets or PII in log lines.
  • Every alert has a runbook link.

Gives 2 of the 12 instructions most monitoring observability skills give in 530 tokens

Counted across 481 of the 483 authors here whose files we hold, read 2026-08-06

  • link every alert to a runbookin 43 of 481, across 35 files
  • use structured json logginghere, and in 36 of 481, across 31 files
  • alert on user-facing symptomsin 20 of 481, across 15 files
  • emit structured JSON logs with stable event namesin 18 of 481, across 13 files
  • propagate trace context across boundarieshere, and in 16 of 481
  • use histograms for latency trackingin 14 of 481, across 9 files
  • use OpenTelemetry for distributed tracingin 13 of 481, across 8 files
  • include a correlation ID on every log linein 13 of 481, across 8 files
  • Define service level objectivesin 10 of 481, across 7 files
  • Call useAzureMonitor before importing other modulesin 9 of 481, across 2 files
  • stop and ask for clarification if inputs are missingin 9 of 481, across 2 files
  • define on-call questions before adding telemetryin 9 of 481, across 4 files

Said here and by no other author read

  • instrument first before debugging
  • log at operation boundaries
  • instrument workers with the USE method

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.