1 big Google Cloud SQL instance, 2 small Google Cloud SQL instances or 1 medium + 1 replica?

Question

I've started to use Google Cloud SQL and I need to improve my IOPS and network speed. I've seen that this it's only possible improving the type of machine and/or improving the size of disk. And this is my question. In my case, I need to migrate 2 MySQL databases (from 2 different projects) and I don't know what is better: 1 big instance with 2 databases? 2 small instances with the database in each instance? or 1 regular instance + 1 read replica instance?

Thank you in advance!

Answer 1

The answer is usual "It depends".

If you're not concerned with data isolation issues, a single instance would be more efficient and easier to manage.

If you split data between instances, you're also capping performance per database. This can be a non-issue if your datasets are similar and process the same amount of requests.

Read replicas could be a solution to scale IOPS if your application workload is heavily skewed towards reads.

Also, independent of which option you will choose, consider HA-setup.