- 최신
- 최다 투표
- 가장 많은 댓글
Aurora Serverless v2 is an on-demand, autoscaling configuration for Amazon Aurora. Check this documention:
Yes it's good for a production transactional system, particularly if your workload is somewhat spiky. Your cluster can be a mix of provisioned & serverless, so the provisioned part can cover your base load and serverless can cover the spikes. This is better than covering it all with serverless v2 as it can get expensive.
Compared to serverless v1 is has faster scale up & down and will scale in smaller increments. Scaling is non-disruptive - no dropped connections.
One thing it's currently missing for infrequent-use workloads is that it doesn't scale down to 0, so you could argue it's not truly serverless!
Short answer is Yes. If system that you are planning run is having spiky or non-consistance load, Serverless V2 is perfect fit. V2 supports almost all the feature available with Aurora Postgres.
This can help in saving lots of cost and almost zero efforts in allocating compute for peak usages as it will grow when it is required.
If you are certain about the usages of your system, or plan for some Proof of concept or Performance test before going live on it to validate that system that you are planning use for V2.
also worth verify constraints:
https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/CHAP_Limits.html
