Installer thanos avec Homebrew, apk, Nix

Liens du projet

• Page d'accueil
• Dépôt
• Docs amont
• Page du gestionnaire de paquets

Commandes et alias

• thanos

Historique du projet

Thanos was created in the Prometheus ecosystem to address operational gaps that appear when teams run many Prometheus servers or need longer metric retention than local Prometheus storage comfortably provides. The CNCF incubation announcement says the project was created at Improbable to enable reliable monitoring on a global scale.

The project architecture follows a component model rather than a single monolith. Its README lists aims including a global query view, unlimited retention, and high availability; its getting-started documentation presents Thanos as a single binary whose features can be deployed independently for gradual rollout.

Thanos moved into CNCF governance as the cloud-native monitoring ecosystem matured. The CNCF project page records acceptance into CNCF on July 14, 2019 and movement to Incubating maturity on August 19, 2020.

Historique d'adoption

The official Thanos repository maintains an adopters list including organizations such as Aiven, Monzo, Adobe, eBay, Sauce Labs, Tencent, Wise, ByteDance, SoundCloud, and many others. This gives the project unusually visible enterprise adoption compared with small CLI-only formulae.

The Thanos documentation also records talks, user stories, and blog posts from 2018 onward, including an introductory 2018 post, 2019 user stories, and later KubeCon/CNCF-era talks. That official documentation trail shows the project moving from early Prometheus-at-scale experiments into a standard option for multi-cluster and long-retention metrics systems.

Modes d'utilisation

Operators use the `thanos` binary to deploy components such as Sidecar, Query, Store, Receive, Rule, Compact, and Query Frontend. The getting-started guide explains that Thanos can be added gradually and can run in Kubernetes or more traditional environments.

The package is especially relevant to Prometheus users who need object-storage-backed historical metrics, deduplication across high-availability Prometheus pairs, global query views across clusters, and downsampling for older data.

Pourquoi les passionnés de paquets s'y intéressent

For package nerds, Thanos is a case where a Homebrew formula is not just a developer convenience but a way to get the same operational binary used in Kubernetes and cloud-native production systems. The README notes that release tarballs are built for major platforms and that every commit to main builds container images.

Its package-manager footprint in the input spans Homebrew, Alpine, and Nix, while its primary distribution story remains GitHub releases and OCI images. That split reflects modern infrastructure tooling: package managers help local testing and ops workflows, while production often consumes containers or Kubernetes manifests.

Chronologie

• 2018: Official docs list the introductory Thanos blog post and first Prometheus meetup slides.
• 2019: CNCF accepts Thanos on July 14, according to the CNCF project page.
• 2020: CNCF moves Thanos to Incubating on August 19.
• 2024: The README notes ThanosCon as a KubeCon Europe co-located event.

Related projects

• Thanos is directly tied to Prometheus and its storage format, and the README positions it as an add-on to existing Prometheus deployments.
• The getting-started documentation points to prometheus-operator, kube-thanos, and community Helm charts as common Kubernetes deployment paths.

Sources

• CNCF incubation announcement: https://www.cncf.io/blog/2020/08/19/toc-approves-thanos-from-sandbox-to-incubation/
• CNCF project page: https://www.cncf.io/projects/thanos/
• Official README: https://github.com/thanos-io/thanos#readme
• Official adopters list: https://github.com/thanos-io/thanos/blob/main/website/data/adopters.yml
• Official docs: https://thanos.io/tip/thanos/getting-started.md