Alerting
Questions about alerting and problem detection in Dynatrace.
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Where I can check my custom events created from anomaly detection using dql over logs data

heramb_sawant
Organizer

Hi,
I  have created anomaly detection to capture specific error from the logs.In that I have defined dql that get timeseries data for the error occurrence for particular k8s application. Statis threshold is defined.

In the anomaly list , status shows "success". I verified target error has occurred , seen in the logs as well.But somehow  I am not able to see my custom event in in grail-> events.

Can somebody please help me to find all the custom events in grail which are related to all the Anomaly detection.


Note - My final goal is to trigger workflow if custom events occurs.

Any help will be appreciated.


Regards,
Heramb Sawant




8 REPLIES 8

heramb_sawant
Organizer

Any help would be appreciated.

Hi,
dt.system.events will usually not show the actual custom event raised by anomaly detection. That bucket is mainly for detector execution/status information, such as ANOMALY_DETECTOR_STATUS_EVENT and ANALYZER_EXECUTION_EVENT—for example, whether the detector ran successfully, timed out, or produced warnings
If the anomaly detection configuration really raised an alert, I’d look for it in Davis events, for example:

fetch events, from: now()-24h
| filter event.kind == "DAVIS_EVENT"
| sort timestamp desc

events.jpg

and then narrow it down using your event template fields / event name if available. Dynatrace treats events from custom alerts as Davis events, which can then be correlated into Problems

So I’d separate it like this:

  • dt.system.events → anomaly detector execution health/status
  • events / Davis events → actual alert events raised by the detector
  • Problems → correlated incidents created from those events

System event models 
For example, if you want to see the related active problems created from custom alerts:

fetch dt.davis.problems, from: now()-24h
| filter event.status == "ACTIVE"
| filter event.category == "CUSTOM_ALERT"
| fields timestamp, display_id, event.name, event.description, event.category, event.status, affected_entity_ids
| sort timestamp desc

custom_alert.jpg
So if your final goal is to trigger a workflow, I’d first verify whether the anomaly detector is producing a `CUSTOM_ALERT` Davis event, and then decide whether to trigger on the event itself or on the correlated problem.

AntonPineiro
DynaMight Guru
DynaMight Guru

Hi,

Maybe you can check System event models.

Best regards

❤️ Emacs ❤️ Vim ❤️ Bash ❤️ Perl

Thanks t_pawlak, you shared very useful information and thanks to AntonPineiro


I referred System event models. ,  checked  below DQL, and found my task is failed with reason "The task was disabled by the service operator to maintain system stability. Contact support if the task keeps failing."

fetch dt.system.events, from:now()-24h
| filter event.kind == "ANALYZER_EXECUTION_EVENT"

what you will suggest  on this?? How it can be enabled.

Note - DQL defined in anomaly detection is  pulling timeseries data capturing specific 4xx error.

Any inputs ??

any further help will be appreciated 

Hi,

I tested the same DQL on my side:

fetch dt.system.events, from:now()-24h
| filter event.kind == "ANALYZER_EXECUTION_EVENT"

and the query itself works, so this does not look like a DQL syntax issue.

ab.jpg

Based on the error message:

"The task was disabled by the service operator to maintain system stability. Contact support if the task keeps failing."

it looks like the specific anomaly detection task was automatically disabled by Dynatrace safeguards, most likely because the detector execution was considered too expensive or unstable.

So I would suggest review the anomaly detector query and try to reduce its cost, like narrow the timeframe, or use some filters. Make it less expensive. 
And if the issue continues, opening a Dynatrace Support case, because the message suggests a service-side safeguard.

sonja
Dynatrace Champion
Dynatrace Champion

hi! in addition to all the useful comments, I wanted to share this new page in the documentation that helps better understand how everything fit together: https://docs.dynatrace.com/docs/analyze-explore-automate/alerting-and-notifications including this graphic.

I hope this helps,
Sonja
20260316-workflows-problems-2645-abca3e7823


Featured Posts