Hi, I have a situation with a 2M row table using LOG_BASED replication. Occasionally, for unknown external reasons that we can't control, this table undergoes a full update, placing every row back into the log. This can happen multiple times per day. Unfortunately, if this happens whilst Meltano is currently fetching the data, it then has to fetch the new changes before completing, getting itself into a lengthy loop. The specific case aside, is there a generally applicable strategy for handling cases where CDC writes are happening faster than reads? I'd rather be a bit behind than getting into loops, and to full refresh once I'm approaching being behind by the retention period.