Open source
Prometheus
The de-facto standard for cloud-native metrics.
- signals
- Traces
- Metrics
- Logs
- Profiles
- otlp ingest
- native endpoint
- storage
- Local TSDB, remote-write to long-term stores
- website
- prometheus.io ↗
Prometheus pulls metrics by scraping endpoints and stores them in a local time-series database queried with PromQL. Recent releases include an OTLP receiver, so OpenTelemetry SDKs and Collectors can push metrics to it directly.
Best for
Infrastructure and service metrics with alerting.
Strengths
- +PromQL is the lingua franca of metrics
- +Huge exporter ecosystem and Alertmanager
- +Simple single-binary deployment
Consider
- −Single-node by design — use Mimir, Thanos or VictoriaMetrics to scale
- −High-cardinality labels can get expensive fast
Also handles these signals