What would you put on a dashboard for a production API you own?

In production each dashboard has a service owner, review date, linked SLOs, and a clear audience. A standard overview covers user journeys, request health, dependency health, saturation, queues, and deploy markers, while deeper views hold technology-specific detail. Panels link directly to filtered logs, traces, runbooks, and the corresponding alert so context survives a handoff. I verify the dashboard during incidents and game days, track whether panels actually helped mitigation, and remove stale queries after schema or route changes. Access, retention, refresh interval, and high-cardinality costs are part of the dashboard design rather than afterthoughts.