Best practices of using a managed PostgreSQL in the cloud
summary
Managed PostgreSQL in the cloud locks down the OS with no SSH access and replaces root privileges with a limited superuser role while high‑level ops use the provider CLI. Data ingress uses \copy via psql or pg_dump/pg_restore; Schema Conversion Tool is suggested for non‑Postgres migrations. Backups are snapshots plus WAL for point‑in‑time recovery. Common architectures use a local read replica plus cross‑region replicas for DR; tune checkpoint_timeout, wal_compression, hot_standby_feedback and autovacuum to avoid XID wraparound and replica lag.
AI-written summary · about
chapters (19)
- 0:00Intro
- 2:34Deploying PostgreSQL in the Cloud
- 5:49Locked Down PostgreSQL Instance
- 8:36PostgreSQL Superuser
- 11:10Interacting with a Managed Service
- 13:29PostgreSQL Configuration Parameters
- 16:01Deployment Options
- 17:50Better Practices for Production Deployments
- 21:06Host-Based Authentication
- 21:47Virtual Private Network
- 26:57Optimizing PostgreSQL Memory Parameters
- 29:09Avoiding Transaction ID Wraparound
- 31:06AUTOVACUUM Tuning
- 32:31Read Replica Lag
- 37:58PostgreSQL to PostgreSQL Bulk Load Options
- 41:44Snapshot and PITR
- 47:22Application Side Connection Pooling
- 49:47Instance Monitoring
- 50:45PostgreSQL Statistics
related talks
36mPostgreSQL: Ups, DevOps...
42mPostgreSQL Distributed: Architectures & Best practices
38mScaling PostgreSQL Replicas Under Load
46mPostgres Platform "Best Practices" for the Modern DBA
43mPostgreSQL in the cloud: Theory and Practice
30mRunning PostgreSQL on Kubernetes in Production: Day 2 Operations with CloudNativePG
48mMigrating to PostgreSQL: Strategies, Sovereignty, and Success
36mBest Practices in PostgreSQL Tuning: Navigating Key Performance Bottlenecks in the Cloud