Inquiry: Custom Alert Based on Root Cause Pattern in Dynatrace Managed

[Aboud1]

Hello All,

is it possible to configure a custom alert in Dynatrace Managed based on a specific pattern appearing in the root cause section of a problem.

We are trying to build an alerting rule that detects certain exception messages or specific error text patterns that appear under the Problem > Root cause section.

---

Example (for illustration only):

it shows a case where the root cause includes the following pattern:

Error while creating customer interaction record for product:XXXXX

And the related exception is:

java.lang.IllegalArgumentException

We are not trying to create an alert for this specific case, but it is a good example of the type of root cause pattern we would like to monitor and alert on.

We want to know if there is any method or workaround that allows us to:

1. Alert based on specific keywords or patterns in the Problem root cause section.

2. Create custom events if a certain exception type or message text is detected.

---

Environment:

• Dynatrace Managed

• Logs ingestion: ❌ Not enabled/preserved

---

Looking forward to your guidance.

[radek_jasinski]

Hi,

In Dynatrace Managed, it’s not possible to create alerts directly based on text found in the “root cause” section of a problem. However, a similar outcome can be achieved by configuring log monitoring and creating custom events based on specific phrases or exceptions, which can then trigger alerts.

Radek

[Aboud1 · Accepted solution]

Hi Radek,

Thanks a lot for your input 🙏. I managed to find a solution that works without relying on log monitoring. Let me share the steps in case it helps others:

Steps I followed:

1. Create a Request Attribute

   • Navigate to Settings → Service-side service monitoring → Request attributes.

   • Add a new request attribute that captures the exception class or error message from the failing requests.

2. Create a Calculated Service Metric

   • Go to Settings → Server-side service monitoring → Calculated service metrics.

   • Create a new metric with:

     • Metric source = the request attribute you created

     • Unit = Count

     • Optional: Add a filter condition so it only counts when the error message contains your specific phrase.

3. Create a Custom Event for Alerting

   • Navigate to Settings → Anomaly detection → Custom events for alerting.

   • Create a new event with:

     • Metric = the calculated service metric

     • Threshold = ≥ 1 (so if the error shows up at least once, it triggers an alert)

     • Assign the right management zone so the correct team gets the alert.

Result:
Now, even if Dynatrace doesn’t automatically generate a problem card, an alert will still be raised whenever this specific error message appears.

Hopefully this will be useful for others facing the same scenario ✌️