RDS Blue/Green deployment stuck in PROVISIONING for 36+ hours (PostgreSQL 12.22, ap-south-1)

Based on the information provided, your RDS Blue/Green deployment being stuck in PROVISIONING for over 36+ hours is unusual and likely indicates an issue that needs investigation.

While you've mentioned that you've already frozen DDL commands, which is a good practice, there are several potential causes for a Blue/Green deployment to stall:

1. Replication issues: For PostgreSQL Blue/Green deployments, logical replication is used between environments. If there are large objects or certain DDL changes that can't be replicated, this can cause problems. Even with DDL commands frozen, there might be pre-existing conditions affecting replication.

2. Internal replication slots: Sometimes internal replication slots can cause issues with Blue/Green deployments. These slots (with names like "rds_us_east_1_...") are used during the replication process, and problems with them could potentially stall the deployment.

3. Parameter group configuration: Ensure that your parameter groups are properly configured for the Blue/Green deployment. For PostgreSQL, certain parameters must be set correctly for logical replication to work properly.

4. Regional service issues: There could be underlying service issues in the ap-south-1 region affecting Blue/Green deployments.

Since your deployment has been stuck for over 36 hours, this is well beyond the expected provisioning time. Without Developer Support or higher, your options are limited, but you might try:

1. Deleting the current Blue/Green deployment and creating a new one
2. Checking for any service health issues in your region
3. Verifying that your PostgreSQL configuration supports logical replication properly
4. Ensuring there are no large object modifications in pg_largeobject that might be blocking replication

If these steps don't resolve the issue, you may need to consider upgrading your support plan temporarily to get direct assistance from AWS Support.
Sources
BUG: Aurora PostgreSQL doesn't clean up internal replication slots after Blue/Green deletion | AWS re:Post
Block DDL during logical replication for PostgreSQL | AWS re:Post