Daily CPU spike at 3 AM

ops

(Fran) #1

Hi, we have a daily CPU spike (100%) at 3 AM (aprox) and we don’t know if there are some internal processes that can produce this spike.

We are currently using this version:

  • Version 6.7.4 (build 38452)
  • Community edition

The alerts generated follow this pattern:

CheckCPU TOTAL CRITICAL: total=100.0 user=58.21 nice=33.33 system=3.98 idle=0.0 iowait=0.0 irq=0.0 softirq=4.48 steal=0.0 guest=0.0 guest_nice=0.0
Timestamp: 2018-08-28 03:04:53 +0200
Occurrences: 10

We review some external processes that could be executed at same time, but there is no one at the same time.

Could you give us some information about other internal processes to understand what could be happening?

Thank you so much


(G Ann Campbell) #2

Hi,

The only thing that would explain this on the SonarQube side is if you’re running a really big analysis at that time of day.

Ann


(Fran) #3

Thank you for your reply. The problem appears daily and we check that there is no analysis at same time (big or little). Thanks for the information, if there is no Sonar process, we’ll check another host processes.
Thank you again


(Nicolas Bontoux) #4

Hi there,

I would advise to approach this from a simple monitoring/troubleshooting angle, independently from SonarQube at the top being:

  • use your OS monitoring tools to verify whether it’s actually one of SonarQube’s Java process who is eating up the CPU
  • if you narrow it down to SonarQube, then focus more specifically on JVM monitoring, to understand how processes’ resource consumption whenever you observe abnormal behaviour

java.lang.OutOfMemoryError while starting , disk space usage