Hotspot Question You have an Azure data factory that has two pipelines named PipelineA and PipelineB. PipelineA has four activities as shown in the following exhibit. PipelineB has two activities as…

Question

Exhibits

Answer Area

• An alert notification was sent after the failure of Activity1 in PipelineA.No
• An alert notification was sent after the failure of Activity3 in PipelineA.Yes
• An alert notification was sent after the failure of Activity1 in PipelineB.Yes

Explanation

Azure Data Factory Alert - Failed Pipeline Runs Explanation

Key Concepts to Understand First

The metric is "Failed pipeline runs" - not "Failed activity runs." Each complete pipeline execution that ends in failure counts as 1, regardless of how many activities failed inside it.

The threshold is "Greater than 2" - meaning the alert fires only when there are 3 or more failed pipeline runs within a 5-minute window. Exactly 2 does NOT trigger it.

Statement 1: Alert after Activity1 in PipelineA fails -> No

At the moment Activity1 in PipelineA fails, the total number of failed pipeline runs recorded within the current 5-minute aggregation window has not yet exceeded 2. The threshold condition (Total > 2) is not satisfied, so no alert is sent.

The individual activity failure is recorded, but it only contributes 1 (or fewer, if that run is still in-progress) to the pipeline run failure count - not enough to cross the threshold yet.

Statement 2: Alert after Activity3 in PipelineA fails -> Yes

By the time Activity3 in PipelineA fails, the cumulative count of failed pipeline runs within the 5-minute window has now exceeded 2 (crossed from ≤2 to 3+). This satisfies Total > 2, so the alert fires.

Activity3's failure (or the pipeline run completing as failed) is the event that pushes the running total over the threshold within that evaluation window.

Statement 3: Alert after Activity1 in PipelineB fails -> Yes

Activity1 in PipelineB failing results in PipelineB's run being counted as a failed pipeline run. At this point, the total failed pipeline runs (across both pipelines - the alert covers all pipelines) within the window exceeds 2, triggering the alert.

Note: the alert is configured for both pipelines, so failures from PipelineA and PipelineB are aggregated together in the same metric count.

Memory Tip

"Greater than 2" = the 3rd failure pulls the trigger.

Think of it like a strike system - the alarm only sounds on the third strike, not the second. Also remember: ADF alerts count pipeline run failures, not activity failures - one pipeline run crashing with 4 bad activities still only adds 1 to the count.

No community discussion yet for this question.