does anybody know if events for custom devices are picked up by Davis in root cause analysis? I have developed an ActiveGate extension which creates custom devices with IP addresses and ports. When service monitored by OneAgent on other hosts communicate to this custom device, a service on top of the custom device is created and it shows "(Group name) on port xxx" (that's correct). You can see service flows to custom devices and also custom devices in PurePaths and that's good - relations are ok. See below:
However if I send some problem opening events for the custom device itself (you cannot send to the "service" entity created on top of the custom device from an extension) and the requests from OneAgent monitored service fail because the custom device fails, it does not correlate in the root cause analysis and I still have two problems:
I've tried sending different severities, events from extension code, event based on metrics, ... without luck.
This is the problem with root cause identified as the service on top of custom device. It's based on the opaque web service calls:
This is the service Dynatrace created on the top of custom device:
This is the custom device itself. You can notice there are several problems opened, but none of them was taken into the root cause analysis for the first problem.
I still think Davis should pick up the events on the custom device itself when it detects problems on a service that the custom device provides. Can anyone ask Davis, please?
Solved! Go to Solution.
After reaching out to support to ensure that I get you the correct information, it depends on the custom device itself.
Mainly it depends on how we see custom devices as part of the service flow.
Custom device as proxy -> Not yet part of root cause, but is something that is coming soon.
Custom device as endpoint -> Yes, it is part of the root cause detection
Custom device not in the service flow -> Not going to be used
@Chad T., in this case, it's not a proxy - see the service flow. Traffic ends on the custom device as shown in the screenshots. My question was about events on the custom device itself - as you cannot push any problem opening events on the services of custom devices.
It's similar to standard Oneagent deployment for example if the root cause for response time degradation would be CPU saturation on a host -> CPU saturation on the host entity.
In my case, the failure is caused by the custom device itself, but it seems it is not included in the root cause detection.
We found a merge bug in combination with your scenario which we will fix in release 207. Expect that the merge of Error rate events with error_level custom events on the custom device works once the 207 release goes into production.