Four golden signals (latency, traffic, errors, saturation)
Four metrics from Google’s SRE book that cover most of what to watch on a user-facing service: latency, traffic, errors and saturation.
Latency is how long requests take, measured separately for successes and failures. Traffic is demand, such as requests per second. Errors is the share of failed requests, including wrong answers served with a success code. Saturation is how full the most constrained resource is: CPU, memory, connections or a queue. Similar shortcuts exist: RED (rate, errors, duration) for services and USE (utilization, saturation, errors) for resources. They are a starting point for alerts and SLOs, not a full observability setup.
Related terms
- Observability (o11y): Being able to tell what a running system is doing, and why, from the data it emits: logs, metrics and traces.
- SLO and SLI: An SLI is a measured indicator of service quality; an SLO is the internal target for it, stricter than any SLA you sign.
- MTTD and MTTA (time to detect and acknowledge): MTTD is the average time from a failure starting to someone or something noticing it; MTTA is the time from the alert to a person taking it on.
- OpenTelemetry (OTel): An open, vendor-neutral standard and set of SDKs for producing and sending traces, metrics and logs.
