Notification based on a user action name

[mkulov]

Hi,

We would like to send an email notification to specific users based on a degradation of a specific user action.

Any idea if it is possible to send a notification based on the user action degradation problem for a given user action name?

Thank you,

Martin 

[radek_jasinski]

Hi Martin,

You can create a Key User Action for a specific user action. In the next step, you create a custom event based on the Key User Action and assign it to the alerting profiles.

Radek

[radek_jasinski]

In addition, you can set performance metric for each key user action:

https://www.dynatrace.com/support/help/shortlink/configure-key-user-actions-web

[mkulov]

Hi Radek,

I imagined that I can use the anomaly detection built in problem alerting in order to send email notifications.

What do you suggest looks like the performance metrics for these custom events will not utilize anomaly detection feature in problem detection.

Is this correct?

[radek_jasinski]

Hi,

When using key user actions, you can configure automatic anomaly detection or a fixed threshold for each such action. 

Here is a similar thread: https://community.dynatrace.com/t5/Real-User-Monitoring/Key-User-Action-Anomaly-Detection-Help/m-p/1...

Of course, if you want to send an email informing only that an action has exceed a threshold to a certain group of recipients, you need to configure a metric event for that key action. 

--------

Dynatrace uses three monitoring strategies for such events:

• Static threshold—the threshold that doesn't change over time.
• Auto-adaptive threshold—Dynatrace calculates the threshold automatically and adapts it dynamically to your metric's behavior.
• Seasonal baseline—Dynatrace calculates a confidence band for a metric with seasonal patterns

https://www.dynatrace.com/support/help/shortlink/metric-events

--------

but for metric key queries, only static thresholds are available.

Then you configure a new alert profile and you're done.

Radek

[janusz_dabrowsk · Accepted solution]

Hi Martin,
Here's how I would try to address it, 
1) Create metric for specific user action

2) Create anomaly detection rule based on created metric