AWS::RDS::DBInstance fails to update certain settings if in a VPC shared via RAM

## 1. Title

*AWS::RDS::DBInstance fails to update certain settings if in a VPC shared via RAM*

## 2. Scope of request

Make RDS and Cloudformation work in a shared VPC.
With VPC sharing via RAM Cloudformation can't be relied on when using RDS.

## 3. Expected behavior

- Create a VPC in one account and share via RAM
- Create an Aurora cluster with one instance in the shared VPC in another account.
- Update the instance's `DBInstanceClass` or `EnablePerformanceMetrics` (there are probably more, but these are the two I ran into)
- Stack update should complete successfully instead of events with `InternalFailure`

Trying to work around these limitations by removing the instance from the stack, updating it by hand and then reimporting also fails with `InternalFailure`.

## 4. Suggest specific test cases

Samples:
- Create a VPC in one account and share via RAM
- Create an Aurora cluster with one instance in the shared VPC in another account.
- Update the instance's `DBInstanceClass` or `EnablePerformanceMetrics` (there are probably more, but these are the two I ran into)
- Stack update should complete successfully instead of events with `InternalFailure`

## 5. Helpful Links to speed up research and evaluation

The support mentioned that the team already knows about this.

## 6. Category (required) - Will help with tagging and be easier to find by other users to +1

3. DB (RDS, DynamoDB...)
4. Networking & Content (VPC, Route53, API GW,...)

Start with the issue's shared-VPC reproduction: create the VPC and RAM share, deploy the Aurora cluster and instance, then update DBInstanceClass or EnablePerformanceMetrics. Investigate the resulting CloudFormation InternalFailure and verify that stack updates and the described remove, update, and reimport workflow complete successfully.

Written by the indexing model from the issue text.