PostgreSQL in the cloud: Theory and Practice
summary
Claims that decades of tuning break in the cloud lead to concrete checks: run Bonnie++ on EC2, expect huge variability; network‑attached disks and shared pipes make random reads costly; a nonpartitioned 40M‑row pgbench query varies 45–130s vs 7s when split into eight 50M partitions; EBS vs ephemeral tradeoffs matter; moving a DB to an SSDs + high‑IO instance turns a crippled system into one that can run 18‑month reports.
AI-written summary · about
chapters (6)
- 0:00Start
- 10:57Random Read Cost
- 19:37Recap
- 25:25How Long Will It Take with a Unitary Table and an Index
- 34:21The Io Throughput Level
- 39:29Why Do I Recommend Using Ephemeral Storage for Databases
related talks
36mBest Practices in PostgreSQL Tuning: Navigating Key Performance Bottlenecks in the Cloud
42mPostgreSQL Distributed: Architectures & Best practices
28mRun PostgreSQL in Warp Speed Using NVMe/TCP in the Cloud
44mContinuous Database Administration
46mBig databases, many servers, data center or cloud - it's all here!
31mDemanding the impossible rigorous database benchm
36mAutomated Performance Testing of Postgres
46mUnlock PostgreSQL’s True Performance with Local SSDs