how to troubleshoot: FATAL: canceling authentication due to timeout

hi all,

We are facing a problem with a user login into database. It happens when there is large load and only from time to time.
Once we get this error, the user becomes unusable until database is restarted. (That user is being used by multiple instances of the same application, it also happens using dedicated users for each application, resulting on one of those users being locked out, the rest keep working fine)

The errors is as follows:
LOG: pam_authenticate failed: Authentication failure
FATAL: canceling authentication due to timeout

Our setup:
3 nodes cluster
- Centos 7
- Streaming replication in place (async)
- WAL shipped to an external location
- pooling done at client side
- Centos joined to an Active Directory domain
- Authentication is using PAM module

User is completely fine in AD side since i can use it to login to a standby DB.
I guess there must be a lock that prevents this user to do the first authentication step, but no idea how to find it. I’ve tried common queries to find locks but I can’t see anything relevant.

I would appreciate if someone could point me to the right direction!

Thanks a lot!
Marc.