Operational excellence
summary
Amazon Aurora uses a log‑based storage that shards data into six copies with acknowledgement from four, spans three AZs, and scales to 64 TB with up to 15 read replicas. Failover to a designated read‑replica takes ~30–40 seconds versus 1–2 minutes for EBS RDS patterns; cache warm‑up after failover drops from ~340s to near baseline when clustered cache management is enabled. Use the cluster writer endpoint, trust defaults (75% shared_buffers), test cloning, continuous backups, Secrets Manager rotation, Performance Insights, and query‑plan baselines.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:28RDS portfolio overview
- 2:54Aurora architecture
- 6:17High availability concept
- 12:33Endpoints and failover
- 16:56Caching and performance
- 20:57Memory management
- 26:26Parameter group management
- 31:23Maintenance and cloning
- 37:36Security best practices
- 48:45Backup and restore
- 53:18Vacuum and performance
- 56:34Monitoring and tools
related talks
46mWhat You'll Miss on AWS & How To Find It
37mNavigating Managed Databases for Over a Decade as a Non-DBA
41mResiliency and data residency
26mNavigating Managed Databases for Over a Decade as a Non-DBA
14mExpect the unexpected: Building resilience with AWS
42mA Year with Apache Aurora
38mDeploying Database Clusters in the Cloud
30mArchitecting for Active-Active Resiliency in the Cloud