Citus 11: A look at the Elicorn's Horn
summary
Citus 11 enables query via any node by replicating table schemas, functions, indexes and sequences to workers with automatic schema and metadata syncing, letting clients connect to workers for high-throughput CRUD. New global_pid ties client work to internal shard queries so you can monitor and cancel across nodes. Sequences embed node IDs (so IDs remain unique but not strictly monotonic). Benchmarks on Azure show very high throughput—around 100–200k inserts/sec on multi‑node clusters. Citus 11 Beta and install notes are at aka.ms/citus11beta.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 2:19Citus development
- 6:32Querying via any node (11.0+)
- 7:25Automatic schema & metadata syncing
- 10:26Monitoring a distributed Postgres cluster
- 13:15Configuring for high concurrency
- 16:20Distributed sequences
- 17:32Load-balancing across nodes
- 18:49Measuring insert throughput to Azure
- 21:36Demo of Citus 11
- 25:59Citus 11 beta
- 26:30Q&A with Marco & Claire
related talks
24mCitus from the Customer eyes
08mFoundations of Distributed PostgreSQL with Citus
31mSharding Without the Complexity
16mKeynote – Umur Cubukcu
16mOnline Scale-Out and Cloud Data Ingestion
33mCitus & Patroni: The Key to Scalable and Fault-Tolerant PostgreSQL
31mKEYNOTE: The Distributed PostgreSQL Problem & How Citus Solves it
31mMove Less, Move Faster: Speeding Up Citus Cluster Scaling