Prometheus is the dominant metrics system in cloud-native environments, using an exposition format and pull-based collection: it queries instrumented applications or exporters over HTTP, usually at /metrics. section 4.2
Prometheus also evaluates alert conditions periodically with PromQL. When a rule fires, it sends the alert to Alertmanager, which deduplicates, groups, suppresses, and routes concise notifications to configured receivers. section 7.2.1
From metric to incident
A microservices operator notices an increased error rate on a Grafana dashboard built from Prometheus metrics. An exemplar supplies a trace ID, which leads to Jaeger traces and then service logs, allowing the operator to test whether a recent upgrade caused the problem. section 1.5.2
Use the architecture figure to follow the path from scraped targets through Prometheus and Alertmanager to operator notifications. section 7.2.1
