Navigating Managed Databases for Over a Decade as a Non-DBA
summary
Ten years of running Amazon RDS since 2011 yields concrete, actionable lessons: 100,000 records inserted single threaded shows MySQL on RDS takes ~5m49s versus Aurora ~2m77s in the author’s microbenchmark; Graviton migration caused out of memory crashes when instance memory/CPU ratios differ after aggressive parameter tuning; autoscaling storage and blue‑green swaps wipe Performance Insights history and can orphan backups; provisioned IOPS vs GP3 tradeoffs save or cost thousands; replicas take hours to days to warm, plan accordingly.
AI-written summary · about
chapters (12)
- 0:00Introduction and Agenda
- 1:22The Early Days of RDS
- 6:04Challenges and Solutions
- 8:12Storage Optimization
- 13:21Backup Costs and Optimization
- 15:55Graviton Migration
- 18:31Optimizing Your Instance
- 18:57Migrating from RDS to Aurora
- 20:56Understanding IOPS and Storage Options
- 24:28Blue-Green Deployment Insights
- 29:21Simple Experiment: Aurora vs. RDS
- 34:08Key Takeaways and Final Thoughts
related talks
26mNavigating Managed Databases for Over a Decade as a Non-DBA
01h 05mDatabase management made simple - Amazon RDS (Part 1 and 2)
26mIntroduction to AWS databases: Modernize your data infrastructure
19mManaging Databases at Scale - AWS
28mMove to Managed Databases
21mSave, Grow and Innovate using AWS Databases
40mCloud Databases: Pros and Cons
41mThe Reality of Managing 100+ Managed Databases