Jira increases active DB sessions after DB switchover

[Lukas Konecny]

Hello, our setup is Jira DC(multiple nodes) with PostgreSQL cluster DB. We have issue with high increasement of active DB sessions after DB switchover which stay on the same amount until Jira restart. This is our pool configuration:

<pool-min-size>250</pool-min-size>
<pool-max-size>250</pool-max-size>
<pool-max-wait>30000</pool-max-wait>
<pool-max-idle>80</pool-max-idle>

and connection pool after the switchover:

Anybody has experience? Any parameter to tune the pool so the active sessions(which are broken obviously because we see just some on DB side) are kicked out after some time?

Thank you for each suggestion.

[Sagar]

Hi @Lukas Konecny Welcome to the Atlassian community,

First you need to understand the cause of this database connection pool increased as if you tune/increase the connection pool database resources and not fixed the cause then it will not help even you increase the connection pool to max value,

So, to identify the cause, check the below parameter if they are causing any issue to increase database connection requests and your database is not able to handle those.

• Check the logs for High Traffic and Load from any account in from UI or API
• Are there any Slow Database Queries you can check it in slow query log file in the Jira home directory
• Any bot account or script is run by user causing multiple requests to Jira in short span of time.
• Enable the Jira rate limiting and observe which account is sending large amount of requests to Jira https://confluence.atlassian.com/adminjiraserver/improving-instance-stability-with-rate-limiting-983794911.html

Also, check the related KB document from the Atlassian to get more understanding and update the parameters based on your need.

https://confluence.atlassian.com/jirakb/operations-in-jira-server-fail-with-connection-pool-errors-in-logs-718834030.html

[Lukas Konecny]

Hi @Sagar 

thank you for your response.

The connection pool doesn't have issue at all, you can see the active sessions are around 5 all the time. We are hitting this issue on test instance as well and there is no user traffic basically. We are just interested why the active sessions are increased so rapidly after the DB switchover and are not kicked out after some time?

We have checked the database statistics after the switchover and the DB says there is <5 active sessions.  It looks like there are an orphaned sessions which are not kicked out.

I am going to have a look on it once more again with our DB guys.

Thx.