Golden signals

The golden signals are four measurements, popularised by Google’s SRE book, that tell you most of what you need to know about a service’s health: latency (how long requests take), traffic (how much demand there is), errors (how many requests fail) and saturation (how close to full the resources are). For latency, averages hide the pain, so you look at percentiles: p50 is the typical request, p99 is the slowest 1%. That tail matters more than it sounds, because a page that calls many backends is slowed by whichever one is slowest, so a rare delay on each backend becomes a common delay for users. Watching these four gives you a sensible default dashboard for any service.