We have almost one Terabytes data in Cloud SQL within GCP and we not have any read/write replicas. There is a lot of read/write operations for transactional and analytical workloads.

1. How can we migrate to a more scalable solution(Cloud spanner / Bigquery). We need less downtime and also cost need to be considered

2. We also have looker to be able to get the viaualization of data. How can we consider Looker also in this solution?

3. Please also suggest the options if we want to migrate tho AWS keeping the above two factors along with cost, storage. Also any advantages from if we use partitioning?

CodePudding user response：

I think a lot depends on your project and your implementation. Migrating from Cloud SQL to an other solution, on its own, is not a difficult task, but how are you reading or writing data from your application? Whatever the product you choose to migrate to, the changes you'll need on your application depends on how the application is made.

Apart from this, there are two good guides for your first 2 points:

1. This guide shows you hot to migrate from Cloud SQL (MySQL) to Cloud Spanner. I would not consider BigQuery, because you mentioned that there is a lot of transactional workloads.
2. Looker is now part of GCP. You can find a good guide on how to connect it to Spanner here.
3. If you want to consider AWS, I'd choose Aurora, and here you can find a good migration guide.