Serverless Databases: The Good, the Bad, and the Ugly
summary
Defines serverless relational databases by four requirements and shows how AWS Aurora Serverless meets them while imposing hard limits: MySQL 5.6/5.7 compatible but not MySQL 8, single-instance scaling only, and limited configuration. Scaling uses a 1–256 range (roughly 2 GB to ~0.5 TB RAM), pause/wake can take 20–30 seconds, failover may take up to a minute, no public endpoint or S3 import, and data API trades latency for connectionless access.
AI-written summary · about
chapters (31)
- 0:00Introduction
- 0:33Main characteristics
- 1:22Charging for usage
- 2:08What is serverless
- 2:40What is not
- 4:03Im lazy
- 4:41Elastic application layer
- 5:19AWS
- 9:27Flexibility
- 11:00Scale
- 11:18Outscaling
- 12:15Range
- 12:31Reuse
- 13:07Two Key Aspects
- 13:31Cloud Provider Reality
- 14:02Use Cases
- 14:24Deployment
- 16:00How it works
- 17:02Compatibility
- 17:22Data API
- 18:21Specifications
- 19:13Scaleup
- 19:47Pause
- 20:34Limitation
- 21:10MySQL Limitations
- 22:11Exclusion Time
- 22:37Cool Down Period
- 22:50Scale Out
- 23:28How to Scale
- 24:08No AutoScaling
- 24:20Resources
related talks
50mIntroduction to Serverless Databases for MySQL and PostgreSQL
32mMySQL on the Cloud: a Journey from Virtual Machines to Serverless Deployments
06mWhat Is A Serverless Database?
01h 11mAWS RDS Aurora Serverless, ECS Fargate and Other Serverless Products
43mKeep calm and dive deep into newest database innovations
49mInstant and fine-grained scaling with Amazon Aurora Serverless v2
33mServerless & Databases - your options and best practices
40mCloud Databases: Pros and Cons