Self Hosted Citus using Cloud Computing
summary
SafetyCulture moves a high‑write SaaS workload onto a self‑hosted Citus cluster to handle a >12 TB hottest table with thousands of TPS; tooling and practice include automated AMI builds with Packer+Ansible, infrastructure as code via Terraform, HA via Patroni, backups with pgBackRest, Prometheus monitoring, and synchronous replication for durability. Data migration uses foreign data wrappers and pglogical where possible, bulk loads use FDW-based insert-for-select, several Citus and kernel networking bugs appear and get patched, and production performance and customer experience improve.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:44Is self hosting Citus what you need?
- 4:21Do you have a business case?
- 7:31Tooling choices for self-hosting Postgres
- 10:30Principles for production infrastructure
- 11:19AMI creation
- 13:36PostgreSQL HA
- 19:10Things that caught us out
- 20:42Testing and validation
- 22:21Going live
related talks
28mCitus Cluster: our zero-downtime migration story
52mScaling a Saas application beyond a single Postgres with Citus
32m8 keys to the growth of Citus & Distributed PostgreSQL
16mOnline Scale-Out and Cloud Data Ingestion
32mLoading at Hyperscale: Advanced Data Import Techniques for Citus
44mDistributed Postgres with Citus
30mBuilding geospatial apps on Postgres, PostGIS, & Citus at large scale
13mWhat I learned creating Postgres monitoring metrics for Citus on Azure
