How do I automate the AWS DMS error log deletion to avoid storage-full on my replication instance?