cortex: Open Source Monitoring & Observability

What is cortex?

Cortex is a horizontally scalable, highly available, multi-tenant, long-term storage backend for Prometheus and OpenTelemetry Metrics, aimed at platform and observability teams whose metrics workload has outgrown a single Prometheus server.

What it is

Cortex is a CNCF project written in Go and released under the Apache-2.0 licence. It runs across multiple machines in a cluster rather than on one host, replicating metric data between those machines so that the system stays available when individual nodes fail. A single Cortex cluster can hold data and answer queries for many independent Prometheus sources at once, keeping each tenant isolated from the others, and it writes metric data to long-term object storage instead of to a local disk. It lives in the cloud-native and Kubernetes ecosystem, alongside Prometheus itself and the OpenTelemetry Metrics specification, and its documentation covers architecture, configuration, guides and security on the project site.

The problem it solves is the ceiling that a single Prometheus server hits: one machine has finite throughput and finite local storage, so a growing fleet eventually produces more samples and retains more history than that machine can handle. Cortex removes both limits by spreading ingestion, querying and storage across a cluster, and by pushing retained metric data into object storage buckets. Where an organisation previously ran several disconnected Prometheus instances with no shared long-term store or tenant separation, Cortex gives one cluster that ingests from all of them, isolates their data and queries, and keeps the history retrievable after the original servers are gone.

Key capabilities

Horizontally scalable execution across multiple machines in a cluster, exceeding the throughput and storage of a single machine.
Highly available operation: when run in a cluster, Cortex replicates data between machines.
Multi-tenant isolation of data and queries from multiple independent Prometheus sources inside one cluster.
Long-term metric storage backed by S3, GCS, Swift and Microsoft Azure.
Ingestion of both Prometheus data and OpenTelemetry Metrics, per the README.
Documentation set covering Getting Started, Architecture Overview, Configuration, Guides and Security.
Vendor-neutral governance as a CNCF project, with support through the #cortex channel on CNCF Slack, the [email protected] mailing list, and community calls every four weeks on Thursdays at 1700 UTC.

Who uses it and how

Kubernetes operators who need cluster-wide metrics collection, a workflow reflected in the project's kubernetes topic.
Organisations running several independent Prometheus sources that require isolated data and queries within a single shared cluster.
Teams whose sample volume or retention window has exceeded what one machine can serve, and who therefore need multi-machine scaling.
Large-scale deployments of the kind described in the April 2025 KubeCon talk "Taming 50 Billion Time Series: Operating Global-Scale Prometheus Deployments on Kubernetes".
Contributors and adopters active in the CNCF community through Slack, the mailing list, and the recurring community calls.

Getting started

The README does not inline an install command; it directs readers to the Getting Started page at cortexmetrics.io/docs/getting-started/, with Architecture Overview, Configuration and Guides pages alongside it. The facts provided name no package, Docker image or compose file, so the documented path is the project's own Getting Started guide.

How it compares

No list of paid products that Cortex replaces was provided, and the facts name no peer tools in this registry — Prometheus and OpenTelemetry Metrics are the data sources it stores for, not substitutes for it. On the facts available, it stands alone here.

When to use it — and when not to

Adopting Cortex means operating a multi-machine cluster and maintaining long-term object storage on S3, GCS, Swift or Azure; nothing in the facts indicates a required database or mail service beyond that. Single-server deployments and small teams that fit comfortably inside one Prometheus instance should not take on that operational load. The README is short and largely made up of badges, documentation links and community pointers rather than setup instructions, and the repository carried 343 open issues at the last recorded state — check current issue and release activity before committing.

project readme (upstream, from github) — read inline

Cortex

Cortex is a horizontally scalable, highly available, multi-tenant, long-term storage solution for Prometheus and OpenTelemetry Metrics.

Features

Horizontally scalable: Cortex can run across multiple machines in a cluster, exceeding the throughput and storage of a single machine.
Highly available: When run in a cluster, Cortex can replicate data between machines.
Multi-tenant: Cortex can isolate data and queries from multiple different independent Prometheus sources in a single cluster.
Long-term storage: Cortex supports S3, GCS, Swift and Microsoft Azure for long-term storage of metric data.

Documentation

Community and Support

If you have any questions about Cortex, you can:

Ask a question on the Cortex Slack channel. To invite yourself to the CNCF Slack, visit http://slack.cncf.io/.
File an issue.
Email [email protected].

Your feedback is always welcome.

For security issues see https://github.com/cortexproject/cortex/security/policy

Engage with Our Community

We invite you to participate in the Cortex Community Calls, an exciting opportunity to connect with fellow developers and enthusiasts. These meetings are held every 4 weeks on Thursdays at 1700 UTC, providing a platform for open discussion, collaboration, and knowledge sharing.

Our meeting notes are meticulously documented and can be accessed here, offering a comprehensive overview of the topics discussed and decisions made.

To ensure you never miss a meeting, we've made it easy for you to keep track:

View the Cortex Community Call schedule in your browser here.
Alternatively, download the .ics file here for use with any calendar application or service that supports the iCal format.

Join us in shaping the future of Cortex, and let's build something amazing together!

Is cortex free to use?

cortex is open source under the Apache-2.0 licence. There is no licence fee and no seat count — you can self-host it or, where the project offers one, pay a vendor for a managed version instead.

What does cortex do?

A horizontally scalable, highly available, multi-tenant, long term Prometheus.

What is cortex written in?

cortex is primarily written in Go. Its source is publicly available at https://github.com/cortexproject/cortex, and it has 5,863 GitHub stars.