Thanos
CNCF-graduated long-term storage and global query layer for Prometheus, backed by object storage.
| License | Apache-2.0 |
| Repository | https://github.com/thanos-io/thanos |
| GitHub stars (approx) | 14000 |
| Language | Go |
| Self-host deployment | kubernetes helm, docker, single binary per component |
| Commercial hosted option | no / self-host only |
| Freshness status | Fresh |
| Freshness checked | 2026-05-17 |
| Categories | metrics |
| Replaces SaaS | datadog, new-relic |
Thanos extends Prometheus with object-storage retention, global query across multiple Prometheus clusters, and downsampling for fast historical queries. A small constellation of components (sidecar, store, query, compactor, ruler) bolts on without forcing a rewrite of your scrape config.
Pick it if: you already operate Prometheus at multi-cluster scale and need a global view plus cheap long-term retention on S3/GCS/Azure Blob. The incremental adoption path (start with sidecar + query) is the gentlest in this category.
Skip it if: you’re starting fresh — VictoriaMetrics or Mimir reach similar outcomes with fewer moving parts. Also skip if you do not have object storage in your environment; Thanos without object storage is not Thanos.
Deployment: Helm chart maintained by Bitnami and the community; each component scales independently. Plan capacity for the compactor — it is the surprising bottleneck at scale.