If the storage capacity of a Cloud SQL instance is larger than your application needs, then you can manually reduce, or shrink, your storage capacity to a smaller size.
Storage shrink operations are supported on primary instances only and available in all Cloud SQL editions.
Requirements for storage capacity
The target storage size must be larger than the minimum allowed storage capacity that is determined to be safe for the instance.
For consistent performance, we recommend reserving a buffer of storage capacity for certain operations like re-indexing. We recommend reserving roughly 100 GB or 20% or more of current usage, whichever is larger.
For more information, see Shared responsibility.
Performance
Storage shrink operations might impact instance performance. Cloud SQL uses Compute Engine virtual machines (VMs) with persistent storage disks. The storage provided by Hyperdisk Balanced scales linearly and offers predictable performance.
For more information, see Quotas and fixed limits.
Downtime
Storage shrink operations require instance downtime; instances restart when the operation is complete.
Depending on underlying disk size, storage shrink operations might accrue considerable downtime. If your instance requires limited downtime, rather than using storage shrink capabilities, we recommend migrating your data to a new, smaller instance using Database Migration Service.
Operations
All other operations are unavailable until the storage shrink operation completes.
We recommend that you wait until other operations are finished before running a storage shrink operation. This includes operations such as backing up or importing data.
Storage shrink operations are not supported on shared-core instances. However, you can change your application setup to a non-shared core instance and then perform a storage shrink operation.
Limitations
The following feature limitations apply:
- Storage shrink operations can't be applied to replicas or cascading replicas.
- Not supported with Terraform.