Cortex: Multi-tenant Scalable Prometheus
summary
Cortex stores Prometheus metrics at millions of samples per second scale, adds multi-tenant queries (implemented in v1.7.0) that let Grafana pull multiple tenant IDs via the x-scope-orgid header and merge results, and moves from chunks to block storage on S3 for cheaper, faster reads. Admin features include per-tenant retention, bulk Prometheus block upload, horizontal Alertmanager and shuffle sharding; blocks currently struggle with >2 hour delayed ingest and lack per-tenant sample-rate limits.
AI-written summary · about
chapters (4)
related talks
28mCortex 101: Horizontally Scalable Long Term Storage for Prometheus
37mIntro: Cortex
25mCortex: Ingesting and Querying 100s of Billions of Samples Per Day
39mDeep Dive: Cortex
22mCortex: Ingesting and Querying 100s of Billions of Samples Per Day
20mCortex: Intro and Production Tips
05mMigrating Prometheus data between different storage systems
23mHow we converted our Cortex data to TSDB blocks