Breakout Session
summary
Postgres runs TripAdvisor’s core: over 100 dedicated servers across 4 datacenters, handling spikes beyond 1,500,000 database queries per minute, with cross-site replication and a scheduled role-reversal every 3 months. Real deployments include a sharded table of 2.6 billion tuples, a real-time anomaly DB ingesting 87,000 tuples per second, mixed 9.1–9.3 clusters, SSD-first hardware, and an operations model staffed 1 full-time + 4 part-time — all proven in production.
AI-written summary · about
chapters (18)
- 0:00PostgreSQL at TripAdvisor
- 3:31Infrastructure and operations
- 7:00Site architecture and read workloads
- 10:34Database upgrades and replication
- 12:42Disaster recovery and failovers
- 16:18PostgreSQL at scale lessons
- 20:31Linux optimizations for PostgreSQL
- 25:45Hardware and RAID tuning
- 27:10Network and power for low latency
- 28:11Performance analysis with perf
- 30:18Caching and query optimization
- 33:07Cascading failures and pools
- 35:13Timeouts and read-write separation
- 36:39Hardware specs and bottlenecks
- 38:26Open source tools
- 42:57DNS and BGP rerouting
- 44:47Master failover considerations
- 46:32Q&A on PostgreSQL configs
related talks
46mBig databases, many servers, data center or cloud - it's all here!
13mBuilding reliable infrastructure with Postgres
36mPostgres at any Scale
51mScaling Postgres to the next level at OpenAI
43mJust Use Postgres.
51mIdentity at Scale; How Okta uses Postgres
47mPostgres & Beyond
28mUsing a database in a data science project - Lessons learned in production