We are experiencing Host metrics gaps from OneAgent on AIX ver. 7.2 TL 5 SP 5 PowerPC 64-bit. Observed on Managed with OneAgent v 1.257.250.20230202-090730 and v 1.261.201.20230315-145944. Like CPU, disk space, disk throughput, memory RAM, process availability...
This render the monitoring on theses hosts 100% useless.
Any one experiencing this ? Any idea ?
For the record : already in touch with support :
Solved! Go to Solution.
04 May 2023 11:36 PM - last edited on 09 May 2023 01:58 AM by MaciejNeumann
Hello. Sometimes gaps point that there is network connectivity issues between OneAgent & ActiveGate or Dyntrace Cluster Node.
default log location ruxitagent_host_*:
Example String for search:
|info [comm ] URL https: //220.127.116.11:9999/communication not working (Connection timed out after 2000 milliseconds) (occurred X times in the last 1h 0m 0s)|
In any case, contacting support is a good option.
Totally agree with @Romanenkov_Al3x.
You can also read troubleshoot monitoring interruptions and run OneAgent diagnostics, it will be helpful if you contact Dynatrace ONE/support.
Turns out DB monitoring tool AIX ITM Tivoli GSMA DB2 (among which kuddb2 (/opt/IBM/ITM/aix526/ud/bin/kuddb2) ) seams to be slowing down drastically calls to perfstat_process_util, making OneAgent calls to this utility very slow leading to metrics gaps. I am not sure about what is the root cause kuddb2 or oneagent
We uninstalled the Tivoli thing and tadaa: at once problem gone.
@gilles_tabary there can be any reason for that. If you download the support archive, check the logs for the OneAgent (os agent) and there will be timings about metrics collection. From that you can focus on the area which takes longer than expected and results into monitoring gaps.