We have a pod and want to test the monitoring of the limits of pods that should be running (Max, not min). Try it with the custom alerts under the process (for all instances) but it dont "add up" the number of process running. And selecting the Process Group shows the metrics empy. So no metric could be selected.
Is there any way to do this?
Solved! Go to Solution.
The only idea I have is sending such metrics into dynatrace as custom metric via some script. Then you will be able to set thresholds for it. They will results events + problems later.
I though about it, but it would be simpler to extract from the dynatrace api the PG and how many elements does it have and alert it in our alert system.
What is really odd is the possibility to select the process group but no metrics are listed (in any of them).
You can ping environment api as well to check amount of entities in process group and make alert in your system based on this output. Is it ok for you?
That was my final idea. But the alert wouldn't be in dynatrace sadly.
So pinging environment api May be used to push custom metric to DT with custom threshold and event and in such case you will have alert in DT. You should be able to make simple activegate plugin for this as well.
We've logged a support ticket for the same thing: the Process Group don't have any metrics and selecting a specific process always shows 12 as the sum.
Hi @Andre van der V. Yes, i notice that. for me it was 6 all the time. I actually managed to make it kinda work, but it forced me to "Strip" the concept of docker at process level and unite all the pods under a same process (no pod name or id).
In that case i notice that the Pods running was actually the detected Workers for that process, 3 pods = 3 workers and so on... But that was not use since all the information about the transactions for an specific pod was missing. And also couldnt split the transactions metrics per pods. Either.
Got this back from support:
The "Process process count" metric is the number of PIDs concurrently running per PGI, not the number of PGIs per PG.
Regarding the empty metric list for "Process Group": there aren't any built in metrics on PG level. So unless there are any custom plugins which define custom metrics on PG level, the metric list will be empty, when selecting "Process Group". As this might be confusing, we redesigned this screen (currently being implemented)
Hey @Andre van der V. thanks for the update. Is incredible that this simple Metric can't be actually alerted.