History
Origin
Thanos was started in late 2017 at Improbable in London by Bartłomiej Płotka and Fabian Reinartz, and opened to the public in early 2018. The motivation was a practical operations problem: scaling Prometheus to long retention and a global view without paying for ever-larger local disks. Fabian Reinartz, a Prometheus core developer, and Bartłomiej Płotka described the design and its origin in their conference talk Fabian Reinartz and Bartlomiej Plotka: Thanos.
The name reflects the design: the components combine to give Prometheus powers it does not have alone.
Timeline
| Year | Milestone |
|---|---|
| 2017 | Started at Improbable by Bartłomiej Płotka and Fabian Reinartz. |
| 2018 | Project made public. |
| 2019 | Accepted into the CNCF as a Sandbox project (2019-07-14). |
| 2020 | Promoted to CNCF Incubating (2020-08-19). |
| 2026 | Latest stable release v0.41.0 (2026-02-12); v0.42.0-rc.0 cut 2026-06-23. |
How it evolved
Thanos began as a sidecar-plus-object-storage extension to existing Prometheus servers: the sidecar uploads TSDB blocks and exposes a StoreAPI, the store gateway serves those blocks back, and the querier merges everything. A push-based ingestion path, the Receive component, was added later so that workloads that cannot be scraped can remote-write into Thanos. Both ingestion styles are now supported side by side, which the recon notes describe as the project's two front doors.
Governance moved away from a single company toward CNCF multi-vendor stewardship. The TOC incubation announcement marks that step, and the core maintainers listed in MAINTAINERS.md now span several employers including Google, Polar Signals, Vinted, Red Hat, AWS, Shopify, and Cloudflare.
Where it stands now
Thanos is a CNCF Incubating project (see the CNCF project page) under multi-vendor maintenance. Releases land roughly every six weeks as a single binary on GitHub Releases plus a quay.io/thanos/thanos container image. A CNCF TAG Security self-assessment documents its security posture.