I'm currently managing a PostgreSQL flexible server database in the East US region, and we've recently faced a significant increase in traffic. We attempted to scale up the database, but Microsoft turned down our request due to capacity issues in the East US. We're considering the option of migrating to another region to accommodate our needs. What would be the best approach to minimize downtime during this migration process? Most of our other services are in East US, and I'm concerned that bandwidth might be affected if we move the database elsewhere. Has anyone handled similar situations?
2 Answers
It seems like Azure has some capacity challenges across different regions, especially in East US. I've had to deal with support endlessly for just a few additional cores! You're not alone in this struggle.
A good strategy would be to set up a read-only replica in the new region using a higher SKU. Once that replica is ready, you can stop writing to the original database and enter maintenance mode for your applications. After that, switch over to the replica for writes. The failover process is really quick, typically around 5 minutes. If you're using a custom domain for your database connection, make sure to lower the TTL. This will help make the transition smoother and prevent any writes trying to hit the old database.
That sounds promising! Just to clarify, will there be added egress charges since the app in East US will be accessing the database in another region?
Yeah, I’ve noticed East US has been particularly tight on resources.