Prometheus

Prometheus is an open-source monitoring and alerting system originally developed at SoundCloud in 2012 by Julius Volz and Bjorn Rabenstein, inspired by Google's internal Borgmon monitoring system, and now a graduated project of the Cloud Native Computing Foundation widely adopted across the cloud-native ecosystem as the de facto standard for metrics collection in Kubernetes environments. The multi-dimensional data model stores time series identified by metric names and key-value label pairs, enabling flexible querying and aggregation across any dimension, with four metric types including counter for monotonically increasing values, gauge for arbitrary values, histogram for distribution sampling, and summary for quantile calculations. The PromQL query language provides powerful aggregation, arithmetic, and statistical operations over time series data, supporting rate calculations, histogram quantile estimation, prediction with linear regression, join operations, and subquery support, enabling sophisticated analysis of monitoring data directly within the database. The pull-based collection model scrapes metrics from instrumented applications through HTTP endpoints, with service discovery automatically finding scrape targets in Kubernetes, EC2, Consul, DNS, and other environments, while the Pushgateway handles short-lived batch jobs. The alerting manager handles alert routing, grouping, inhibition, and silencing with notification delivery to email, Slack, PagerDuty, and webhook integrations. The local storage engine uses an efficient custom time series database with downsampled recording rules for long-term retention. Adopted by Uber, DigitalOcean, GitLab, and thousands of organizations. Designed for site reliability engineers and platform teams.