10.1.3 · Skill in Product observability
Service metrics
Numbers that describe system health, such as error rate and latency.
Service metrics track rates, errors and durations over time. The RED and USE methods give simple starting sets. They are cheap to store and the basis for dashboards and alerts.
Resources
1- bookStart hereMonitoring distributed systems (Site Reliability Engineering, chapter 6) ↗Google SRE · 20 min
The four golden signals and the case for alerting on symptoms users feel rather than on every cause.
Tools
1- ToolOpen sourcePrometheus
Open-source metrics database and alerting, the common default for service metrics.