Logical Replication Delay on Remote Server

Hi Team,

adding one more point here,

5. Pg_stat_replication_slot total_txn count reduces when lag occurs .Does
decoding stop in case of logical replication delay?

On Tue, 23 Apr 2024 at 22:55, Ramakrishna m <ram.pgdb@gmail.com> wrote:

Hi Team,

I am facing replication lag in postgres16 at times,not able to find the
reason.

Please find the configuration:
1) Two replication slots for two servers (1 is the same data center,
another remote)
2) weekly once/twice facing lag(around 2GB for 1k TPS environment) in the
remote replication slot and another same data center server slot with 0
lag.

My observation:

1) Noticed pg_stat_replication_slot total_txn reduced from 1k to 5 or 6
but other slot is same with 1k TPS
2)And noticed lag for remote slot but fine with same data center
3) Most importantly I have plenty of Bandwidth available in n/w ,2GB is
still freely available out of 4GB network.
4)No IO issues on servers
I am not able to prove if this is due to Network .Can you help me how to
proceed on this?
At Least how the logical decoding total_txn is counting the view
pg_stat_replication_slot?

--
Thanks & Regards,
Ram.

--
Thanks & Regards,
Ram.