2011-03-23 03:47 AM
|About DataFabric® Manager server|
I'm having trouble setting up the alarm for total_processor_busy as well as cpu_busy. Any help?
What I do is set a threshold for system:total_processor_busy and system:cpu_busy and set threshold like 50% with 60 and 300 seconds interval.
Then I set the alarm to send email to me on perf:system:total_processor_busy:breached as well as perf:system:cpu_busy:breached and no email received regardless of how i tweaked my threshold.
Anyone has any idea? Perhaps I'm doing this wrong.
2011-03-23 04:06 AM
Is the event getting generated.?
Can you get the events list using the cli dfm report view events-perf ?
Are you seeing any error in the alert.log in the dfm install dir under log folder ?
2011-03-23 05:02 AM
You may cross check the value of these counters using 'dfm perf data retrieve' CLI for the particular counter.
Check the value of the counter to see if it has exceeded the threshold.
Also paste the output of the 'dfm perf threshold list -x' CLI as well.
Please note that the event would be marked history if it has breached and back to normal. For e.g: Letz say the threshold value was created at time t0, the threshold value is 50% and threshold interval is ti secs. At time t1, if the value of the counter is 51 % (consistently above 50% for the threshold interval i.e., t1- ti time duration), then perf event would be generated. However, at time t2 this event would be marked history (and hence NMC won't show in the console) if the value of the counter falls below 50%. Please verify if this condition applies to your case.