Absorbing Thanos Infinite Powers for Multi-Cluster Telemetry
Talkat KubeCon + CloudNativeCon North America 2020 Virtual · on Thursday, 2020-11-19· at 15:45· Intrado Virtual Event Platform,·
summary
Thanos turns Prometheus into a global, highly available monitoring system by adding sidecar upload to object storage (any S3‑compatible), a querier that deduplicates across Prometheus instances, a store component reading blocks from object storage, a compactor that does downsampling and block optimization, and a receiver that accepts Prometheus remote_write. Recent additions include concurrent select execution, metadata/chunk caching via memcached, and a query-frontend that cuts a 5+ second range query to ~1s in the demo. Multi‑tenant routing uses prom-label-proxy plus an auth proxy to inject tenant labels.
AI-written summary · about
chapters (15)
- 0:00Intro
- 0:15Who are we
- 0:50Overview
- 4:44Components
- 5:40Usage
- 7:25Thanos Courier
- 9:20Layers
- 10:27Longterm Storage
- 14:30Receive Component
- 16:32Conclusion
- 17:28News
- 20:45Demo
- 25:19Wrapup
- 28:10Demonstration
- 31:51Demo with Katakura
related talks
26mPrometheus That Scales
27mThanos: Highly Available, Pluggable, Long Term Metric Storage for Everyone!
36mRunning the Observability As a Service For Your Teams With Thanos
10mUpstream-First, High Scale Prometheus Ecosystem
22mMonitoring the World: Scaling Thanos in Dynamic Prometheus Environments
25mMulti-Cluster Observability: What you need to survive with a fleet
21mMultiverse of Thanos: Making Thanos Multi-Tenanted
31mHandling Billions of Metrics with Prometheus and Thanos