Hi all,
I have been working on migrating ITM situations into Dynatrace, and one of the scenarios involves tracking process memory consumption and generating alerts when usage exceeds a defined threshold.
To replicate this alerting mechanism in Dynatrace, I was referred to creating a metric event.
I have defined the following metric selector for this purpose:
builtin:tech.generic.mem.workingSetSize:filter(and(or(in("dt.entity.process_group_instance",entitySelector("type(process_group_instance),entityName(~"proc_name~")"))))):splitBy("dt.entity.process_group_instance"):auto
This query works as expected for triggering the alert based on process memory usage. However, the problem notification generated only includes Process Group data, making it difficult to determine which host this process is running on.
Issues I Encountered:
- I explored the available dimensions for the metric and attempted to modify the event template, but I couldn’t find a way to attach the host name to the event.
- This is not the first time I’ve encountered this issue—some metrics in Dynatrace do not natively include the host name as part of the metadata.
My Questions:
- Is there a way to retrieve and include the related host entity for a process in the metric event? Alternatively, is there another metric that provides process memory usage while also including host-level information?
- Is there a global approach to associating the host name with every related metric event? (For example, can this be achieved using tags or another built-in Dynatrace mechanism?)
I appreciate any insights or best practices on this.
Thanks in advance,
Roei