PostgreSQL in Kubernetes: Operation - Day 2
summary
Operator code written in Go runs as a deployment (no StatefulSets) and manages PostgreSQL clusters via Cloud Native PG and an EDB fork, claiming doesn't use StatefulSets. Demos show automatic failover, TLS-encrypted replication, pgbouncer operator, Prometheus exporter and Grafana dashboards, rolling minor upgrades (15.3 → 15.4) and pgbench tests. Backups use S3 or Kubernetes volume snapshots — a customer example moves from 4 TB to 8 TB and reports 8 TB recovery in 2 minutes after initial 4-hour first backup and cites 25 million downloads for the operator.
AI-written summary · about
chapters (11)
- 0:00PostgreSQL in Kubernetes intro
- 2:57Why use a PostgreSQL operator
- 7:55EDB operator features
- 16:47Storage for PostgreSQL in K8s
- 20:22Rolling upgrades and backups
- 26:29Live demo setup
- 33:06Monitoring with Prometheus
- 36:08Failover and switchovers
- 40:14Minor version upgrades
- 43:30Backup and restore workflows
- 44:33Q&A
related talks
34mRun PostgreSQL:The Kubernetes way
50mDemystifying Kubernetes for Postgres DBAs; A Guide to Operators
26mKubernetes from the Database Out
30mRunning PostgreSQL on Kubernetes in Production: Day 2 Operations with CloudNativePG
28mLightning Talk: Cloud Native PostgreSQL - Running PostgreSQL on Kubernetes
27mFrom Zero to Hero: Scaling Postgres in Kubernetes Using the Power of...
25mPostgres on Kubernetes for the Reluctant DBA
37mData On Kubernetes, Deploying And Running PostgreSQL And Patterns For Databases In a Kubernetes Cluster.