Prometheus

Open-source systems monitoring and alerting toolkit that collects and stores time-series metrics via a pull-based model, purpose-built for reliability and cloud-native, microservices-based architectures.

Home Page

https://prometheus.io/

Key Features

1. Multi-dimensional data model with metrics identified by name and key/value label pairs
2. PromQL flexible query language for aggregation, filtering, and analysis
3. Pull-based metric collection over HTTP with configurable scrape intervals
4. Built-in Alertmanager for alert routing, deduplication, grouping, and notification (Slack, PagerDuty, email)
5. Automatic service discovery for Kubernetes, Consul, EC2, and more
6. Autonomous single-server nodes with local time-series database (no distributed storage dependency)
7. Push support via Pushgateway for short-lived jobs
8. 100s of community exporters (Node, Blackbox, MySQL, Redis, etc.)
9. Native Kubernetes and container-level monitoring
10. Pairs with Grafana as the de facto visualization layer

Background

1. Originally built at SoundCloud in 2012; second CNCF graduated project (after Kubernetes).
2. De facto standard for Kubernetes and cloud-native metrics monitoring.
3. Standalone, reliability-first design — works even when the rest of your infrastructure is down.
4. Commonly deployed as part of the Prometheus + Grafana + Alertmanager stack.