Cortex: Intro and Production Tips
Talkat KubeCon + CloudNativeCon North America 2021 · on Friday, 2021-10-15· at 11:00· Concourse Hall 153 ABC + Online·
summary
Cortex is a Prometheus-compatible, multi-tenant time-series store that runs across dozens of servers and handles millions of samples per second, using S3-style blob storage for durability and a KV store like etcd/consul for automatic sharding. Core components are distributor → ingesters (two‑hour blocks) → compactor (creates 12/24‑hour blocks), plus query‑frontend, queriers and store‑gateway with heavy caching. AWS AMP runs Cortex and contributes features like zone aware replication, SMS Alertmanager delivery, block‑storage series deletion and has helped fix two CVEs; the project records 293 contributions across recent releases.
AI-written summary · about
chapters (17)
- 0:00Intro
- 0:39Alvin Lin
- 0:57Agenda
- 1:32What is Cortex
- 2:26Who uses Cortex
- 2:58What Cortex does
- 6:16The query path
- 8:18Project update
- 9:00Highlights
- 9:28Distributed Tracing
- 10:51Features
- 11:32Improvements
- 13:29Block Storage
- 13:56Customer Feedback
- 14:52Vulnerabilities
- 16:00Project Updates
- 17:26Cortex vs Thanos
related talks
37mIntro: Cortex
39mDeep Dive: Cortex
28mCortex 101: Horizontally Scalable Long Term Storage for Prometheus
25mCortex: Ingesting and Querying 100s of Billions of Samples Per Day
22mCortex: Ingesting and Querying 100s of Billions of Samples Per Day
30mCortex: Insights, U...
16mCortex: Multi-tenant Scalable Prometheus
45mScaling Prometheus with Cortex, one time series at a time