ec2-gc: DeleteSnapshotInMap is failing

It seems we must delete a snapshot only after the corresponding ami is released.

> {
"Error": "Ec2.Ec2Exception",
"Cause": "The snapshot snap-0xxxxxxxxxxxx is currently in use by ami-xxxxxxxxxxxx (Service: Ec2, Status Code: 400, Request ID: 25aa2) (SDK Attempt Count: 1)"
}


Start at the DeleteSnapshotInMap entry point and trace how the AMI and snapshot cleanup requests are ordered. Reproduce the EC2 error if possible, then verify that the AMI is released before its corresponding snapshot is deleted and that the reported in-use failure no longer occurs.

Written by the indexing model from the issue text.