RED Method
A minimal per-service dashboard: Rate, Errors, Duration — the request-centric view of whether users are being served.
Three measurements per service: Rate (requests per second), Errors (failed requests per second, or the failure ratio), Duration (latency distribution, in percentiles).
Its virtue is that it is uniform. Every request-driven service gets the same three panels, so an engineer can triage a service they have never seen. It maps directly to SLIs, and it describes the user's experience rather than the machine's condition — a host at 90% CPU serving every request in 50 ms is not a problem, and an alert on the CPU would say otherwise.
Pair it with the USE method for resources: for every resource, Utilisation, Saturation and Errors. RED tells you the service is degraded; USE tells you which resource ran out. Between them they cover most first-line diagnosis.