observability

Plugin observability in Adminhead/claude-plugins. Commands read from the repo's manifest — today.

Skills (4)

• dashboards

  Design and build Grafana dashboards from operator questions (is it healthy, where is the bottleneck, what changed, what should someone do) rather than from the list of available metrics, then keep them as code. Covers RED and USE panel sets, a quality checklist, current panel types, datasource UIDs,

• instrumentation

  Add and review production observability in application code, covering structured logs, metrics, OpenTelemetry traces, profiles, and alerts, plus SLIs, SLOs, and burn-rate alerts. Language-agnostic with short Go, Python, and TypeScript snippets. Use when instrumenting a service, adding spans or metri

• postmortem

  Write a blameless postmortem after an outage or significant degradation, covering metadata, a timeline with gap flags, 5 Whys to a fixable system gap, root cause versus contributing factors, impact, detection gap, where we got lucky, and owned and dated actions. Includes a short template for minor i

• promql

  Query and debug Prometheus-compatible backends (Prometheus, Thanos, Mimir, VictoriaMetrics) and write Prometheus server config. Covers PromQL rules for counters, histograms, and query cost, USE and RED investigations, data gaps via `up`, scrape configs, recording and alerting rules, Alertmanager rou