Instance size of a read replica in secondary region of Aurora PostgreSQL Global Database
When creating a read replica in a secondary region of Aurora PostgreSQL Global Database, if possible, we want to use the minimum instance size as much as possible to minimize our running cost because our use case of the read replica is just for DR under the pilot light strategy. But, we're concerned about the replication lag when we use smaller instance size for read replica in the secondary region than that of primary cluster. Is there any risk of replication lag, or any other risk when taking such architecture?
- Language
- English
- Newest
- Most votes
- Most comments
Aurora Global Database uses a physical replication method at the storage layer, so replication lag is not affected by replica instance size. If you want to reduce costs, you can use a headless Aurora DB cluster.
However, there is an ICE concern with headless Aurora DB clusters when the entire region fails over. To solve this, we could run it with Serverless V2 beforehand.
