Prometheus

Open-source monitoring and alerting toolkit that collects metrics as time series data. Uses a pull model over HTTP to scrape targets, stores data locally with a multidimensional label model, and provides PromQL for querying and alerting.

Graduated CNCF project since 2018, originally created at SoundCloud. Widely adopted across enterprises for observability stacks. Differs from push-based tools by using a pull model that keeps the monitoring system reliable during outages, since it does not depend on distributed storage or remote services. Pairs naturally with Grafana for visualization and works well with Kubernetes service discovery.

Key features:

• Multi-dimensional data model with time series identified by metric name and key-value labels
• PromQL query language for aggregation, recording rules, and alerting
• Pull-based scraping over HTTP, with push gateway for short-lived jobs
• Service discovery and static configuration for target discovery
• Alertmanager for alert routing, grouping, and silencing
• Client libraries and exporters for instrumenting apps and third-party systems

Teams use it for infrastructure and application monitoring in microservices and containerized environments. Developers track request latency, error rates, and resource usage. SRE teams rely on it during incidents to diagnose problems quickly without depending on downstream services.