During startup of the Debezium Postgresql, we experienced random infinite loop problems in production. It turns out that the underlying issue was the pg_replication_slot_advance taking longer than the database read timeout. Unfortunately, the retry logic causes a different error which masked the root cause. This was tested in 2.2 and 2.4.

https://debezium.zulipchat.com/#narrow/stream/348249-community-postgresql/topic/Inifinite.20Loop.20Restarting/near/422906598

Which use case/requirement will be addressed by the proposed feature?

Postgresql connector startup troubleshooting will be much easier.

Implementation ideas (optional)

Ideally, the warn logging on line 421 of PostgresReplicationConnection.java would include the exception details and stack trace. This would have highlighted the issue right away.