Postgres Platform "Best Practices" for the Modern DBA
summary
Ten concrete practices: own the provisioning via forms + Terraform so clusters are consistent; add measured friction to avoid hundreds of tiny DBs; run a dynamic database metadata service for ownership and fleet queries; use PG bouncer or similar pooler; enable pg_stat_statements and auto_explain with a high initial threshold; do major upgrades via logical replication and cutover by switching a static A/CNAME; require backups, schedule weekly maintenance windows, and prioritize observability over squeaky-low latency.
AI-written summary · about
chapters (11)
- 0:00Introduction and role definition
- 8:34Ten best practices overview
- 11:06Database provisioning process
- 15:16Long-term database health
- 18:06Developer-owned tooling
- 20:32Fleetwide change rollouts
- 22:38Client connection methods
- 24:04Planned downtime importance
- 27:32Observability over latency
- 33:06Automation and direct access
- 39:58Q&A session
related talks
11mThe Seven Principles for Making a PostgreSQL DB Work Fast
30mManaging Postgres at scale: Challenges, Tools & Techniques
05mCommon mistakes of DBAs
18mPower Hour, Managing Databases at Scale – Peter Zaitsev
47mPostgreSQL worst practices
34mDo we really need a DBA to manage our PostgreSQL databases on DBaaS
21mDriving the Future of PostgreSQL Adoption
49mBeyond PostgreSQL 17