Instant and fine-grained scaling with Amazon Aurora Serverless v2
summary
Amazon announces Aurora Serverless v2 that performs in-place scaling non-disruptively, resizing buffer pool (75% of memory) and CPU instantly; capacity unit is ACU (1 ACU = 2 GB memory) with 0.5 ACU increments and a 0.5 ACU minimum, demos show a live cluster jumping from 0.5 to ~10 ACU and admins setting a 64 ACU max. Features include up to 15 readers, global DB across 5 regions with 16 nodes/region, and integration with RDS Proxy and Performance Insights; scaling works and reads follow writer for high‑availability failover.
AI-written summary · about
chapters (17)
- 0:00Intro
- 1:28Aurora architecture supports serverless
- 2:38Database capacity: Cost vs. management
- 3:50Amazon Aurora Serverless v2
- 6:53Scaled out databases
- 9:17Instant, in-place scaling
- 11:42Buffer pool resizing
- 13:32Fine-grained capacity adjustments
- 14:43Scaling factors
- 15:56Configuring capacity
- 23:38Demo: Flash sale
- 26:30High availability and read scaling
- 28:56Multi-AZ high availability
- 34:40Amazon Aurora global database
- 36:51Aurora Serverless global database
- 40:50Amazon RDS performance insights
- 45:03Aurora clusters in mixed configuration
related talks
50mIntroduction to Serverless Databases for MySQL and PostgreSQL
43mKeep calm and dive deep into newest database innovations
55mAmazon Aurora: A deep-dive into what's new
43mDeep dive into Amazon Aurora and its innovations
25mServerless Databases: The Good, the Bad, and the Ugly
28mGet started with Amazon Aurora DSQL
35mAmazon Aurora DSQL: Scalable, Distributed Relational Databases
39mAWS Aurora