cortexproject/cortex

Cortex is a system for storing and querying metrics data at large scale. It is designed to work as a long-term backend for Prometheus, a widely-used tool that collects measurements from servers, applications, and infrastructure, such as CPU usage, request counts, or error rates. Prometheus is typically good at short-term storage on a single machine. Cortex extends it by storing that data reliably across multiple machines and keeping it for much longer periods. The four main properties Cortex is built around are horizontal scalability, high availability, multi-tenancy, and long-term storage. Horizontal scalability means that as data volume grows, you add more machines rather than upgrading a single one, and the work is distributed across them. High availability means data is replicated, so losing one machine does not cause data loss. Multi-tenancy means a single Cortex cluster can accept and separate metric streams from multiple independent teams or customers, keeping their data isolated from one another. For long-term storage, Cortex supports writing data to object storage systems including Amazon S3, Google Cloud Storage, Microsoft Azure, and OpenStack Swift. This keeps costs manageable compared to keeping everything on local disk. Cortex also supports ingesting OpenTelemetry Metrics in addition to Prometheus-format data. The project is a CNCF graduated project, meaning it is governed by the Cloud Native Computing Foundation and has met maturity standards around security, governance, and production readiness. It is regularly featured at KubeCon conferences, with talks available going back to 2016. Community calls are held every four weeks, and support is available through a Slack channel and mailing list.