Dashboards and baselines
List key panels and “normal” ranges; for release observability, pin comparison baselines and environment tags.
Case topic · not in the main 20
Automation & observability
Single-page topic Maps to observability / platform bands
Overview: turn post-deploy health checks, log queries, and metric baselines into repeatable steps for agents in standard, read-only, or controlled environments. Skills should bind data sources and query templates (trace IDs, service names) and declare permission boundaries; threshold judgments must mark when a human must confirm. Canary phases should compare error rates, latency percentiles, and utilization; multi-region deploys should state whether to aggregate per region or globally.
Relationship to the case hub: the closest official categories are Observability & incidents (log clustering, patrols, on-call handoffs) and Platform & release (drift, release comms). This folder is a supplemental explainer without sibling case HTML files; deeper entries live under those categories.
Implementation notes (single page)
On-call and runbooks
Cross-reference on-call manuals so you do not duplicate guidance; parameterize log queries instead of concatenating raw user input.
Trigger and exit (example)
Trigger: deploy finished or alert includes canary. Output: health check results + baseline comparison. Exit: all green or rollback playbook engaged.
Case hub · Observability & incidents Open devops category index Cases overview