I’m trying to set up a custom alert.
My need is to raise an alert when some Java RMI processes are not started in the morning.
We should have a process group with two instances, and on each one, two processes should run.
I was able to set up an alert on processes count as seen on the screenshot below, but I’m not sure it will do what I expect.
At the moment I write this question, we only have one process group instance running (the processes were not started on one server this morning) and no alert is raised (I created the alert afterwards). I also tried to set up an alert on process group instances count, but I can’t, the drop-down list of metrics is empty.
There is a built in problem pattern implemented that you can configure. Just go to the process groups settings screen and configure the minimum number of process that should run.
The metric you chose for your custom alert is the wrong one as it represents the number of processes within a ‘Process Group Instance’, which has a different semantic than what you would like to achieve.
Thanks for you answer. If I understand well, the setting you propose works on the number of process group instances. So I should set it to 2 as I have two hosts.
If I want to be sure that I have got 2 worker processes per process group instance, I should also keep my custom alert in order to be sure that I have my 4 actives processes.