[Fuel CI] Fuel Jenkins is unavailable time to time.

[Andrey Nikitin]

Hello!

Fuel Jenkins -- ci.fuel-infra.org -- is unavailable time to time (April, 13 - ~15.50; April, 14 - ~13.40).
I've found that all cores of CPU were utilized on 100 percent on the server on that time. Looks like, Jenkins was overloaded by any job or another activity.

Please investigate it from application side. We're going to help to find root cause of the problem, if you need our help.

Priority is High due to Fuel Jenkins is a production service and this problem affects many people.

Thanks in advance!

[Aleksandra Fedorova (bookwar)]

There are no unusual activities on Jenkins side

Consider tuning hardware resources

[Igor Shishkin (teran)]

Andrey, please research if have any confirmations for such problems on atop side.
If yes - problem is about SW on the node, there also we could determine the pattern to track
Otherwise - it's mostly possible about networking so will require further investigation.

[Andrey Nikitin (heos)]

As far I can see, that all cores of the CPU were utilized on 100%, no other problems were found. Looks like, that problem occured due to errors in Postfix configuration, which were fixed.

Related messages in the logs:
Apr 13 12:55:44 fuel-jenkins01 postfix/smtp[17589]: connect to ASPMX.L.GOOGLE.com[2a00:1450:400c:c0b::1a]:25: Network is unreachable
Apr 13 12:55:44 fuel-jenkins01 postfix/smtp[17589]: connect to ALT1.ASPMX.L.GOOGLE.com[2a00:1450:4010:c09::1b]:25: Network is unreachable