Today it looks like when the AWS integration encounters an error during a sync it will completely skip the delete phase at the end of the sync. It would be nice to configure the delete phase to still run regardless of if an error was encountered.
Example log: Resync failed with 2. Skipping delete phase due to incomplete state