Spike in AWS RDS Oldest Replication Slot lag

Posted on

Question :

Does anyone know about this “Oldest Replication Slot Lag” in AWS RDS? What does it represent?

I have a PostgreSQL instance on RDS (with Logical Replication switched on) for transferring data from PostgreSQL DB to S3 bucket (in CSV form) using AWS Data Migration Service. The graph of “Oldest Replication Slot Lag” is too high on this instance. and also Transaction Log Disk Space goes up too.

Any help?

Answer :

The lagging size of the replica lagging the most in terms of WAL data received. Applies to PostgreSQL.

If you’re using logical replication you might consider setting up some monitoring around this metric as it could indicate that your logical replication slot on RDS has switched to inactive – meaning it is no longer consuming the transaction logs.

If left unchecked a replication slot in an inactive state can consume all your disk space.

If you’re using DMS, check the task and replication instance metrics to see if anything sticks out as a bottle neck.

Source: https://docs.aws.amazon.com/en_pv/AmazonRDS/latest/UserGuide/MonitoringOverview.html

Leave a Reply

Your email address will not be published. Required fields are marked *