Observability & Monitoring

Claude Code: /plugin marketplace add wshobson/agents → /plugin install observability-monitoring@claude-code-workflows

Four skills that would each be useful alone are arranged so their outputs line up: Prometheus scrape and recording rules feed Grafana panels, and those panels feed SLO views. The SLO skill is the most concrete part, writing service level indicators as PromQL over a twenty-eight day window, availability as the ratio of non-5xx requests to all requests and latency as the share of requests falling under a histogram bucket boundary. That is what turns an error budget into something computable rather than a slogan. Tracing sits alongside with Jaeger and Tempo and a one to ten percent production sampling recommendation, and four agents cover the database, network, performance and observability angles when a template will not do.