Thanos: Highly Available, Pluggable, Long Term Metric Storage for Everyone!
Talkat KubeCon + CloudNativeCon North America 2021 · on Thursday, 2021-10-14· at 17:25· Concourse Hall 153 ABC + Online·
summary
Thanos provides a Prometheus-compatible path to scale and keep metrics for years by running a sidecar next to each Prometheus that uploads TSDB blocks to an S3‑compatible object store, a compactor that merges blocks and performs downsampling for long-range queries, a query component that deduplicates across replicated Prometheus instances and reads historical data via a store gateway, and a store API gateway to serve object-store data — used in production by many companies and incubating in the CNCF since 2017.
AI-written summary · about
chapters (21)
- 0:00Intro
- 0:31About me
- 1:21Prometheus
- 2:46Grafana
- 3:08Alert
- 3:40Retrieval
- 7:35Limitations
- 7:58Crashes
- 9:19Single Point of Failure Downside
- 11:17Unified View
- 11:43Kubernetes Discovery
- 12:38Federation
- 13:23Metric Endpoints
- 14:11Limitations of Prometheus
- 14:39Community
- 16:06Companies
- 16:46Thanos Features
- 18:53Thanos Sidecar
- 19:56Thanos Compactor
- 24:17Cluster Monitor
- 26:13Join us
related talks
39mAbsorbing Thanos Infinite Powers for Multi-Cluster Telemetry
26mPrometheus That Scales
22mMonitoring the World: Scaling Thanos in Dynamic Prometheus Environments
36mRunning the Observability As a Service For Your Teams With Thanos
31mHandling Billions of Metrics with Prometheus and Thanos
33mStory of Correlation: Integrating Thanos Metrics with Observabi...
10mUpstream-First, High Scale Prometheus Ecosystem
23mConnecting Thanos to the Outer Rim via Query API