Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Section
bordertrue
Panel
titleEnvironment
borderStylesolid
  • Dynatrace Portal
  • Alerts
Panel
titleSolution
borderStylesolid

You can configure an alert to be triggered only after a specified number of consecutive errors on a specified number of nodes. Using this feature can prevent "false alarms." A single error or failure might be the result of a random occurrence and isn't always an indication of a performance issue.

You can set the number of consecutive errors, from 1 through 5, for all alert types.

The examples below, for Transaction Failure Alerts, describe how the consecutive errors setting works.

Example 1:

Transaction Failure Alert configuration for a test named Example:

  • Three nodes are configured: node A, node B, and node C.
  • The Number of consecutive errors is set to 3.
  • The Node Thresholds is 1 node.

When any node consecutively reports 3 failures, the alert is triggered.

If node A reports 1 failure and the next run is successful, while the other two nodes run successfully in all test runs, no alert is triggered.

If all three nodes experience 1 failure each, no alert is triggered as none of them meets the criterion of 3 consecutive errors.

Example 2:

Transaction Failure Alert configuration for a test named Example:

  • Three nodes are configured: Node A, Node B, and Node C.
  • The Number of consecutive errors is set to 3.
  • The Node Thresholds is 2 nodes.

With this configuration, an alert is triggered when two nodes have 3 consecutive errors each.

In the first test interval, nodes A, B and C all fail. All nodes have an error count of 1.

In the second test interval, node A runs successfully, node B fails, and node C fails. The error count for node A is updated to 0, node B to 2, and node C to 2.

If in the third test interval, node A fails, node B fails, and node C runs successfully. The error count for node A is updated to 1, node B to 3, and node C to 0. The alert is not triggered because only one node has 3 consecutive errors.

However, if in the third test interval, nodes A, B, and C all fail, the error count for node A is updated to 1, node B to 3, and node C to 3. The alert is triggered because two nodes have 3 consecutive errors.

 Node AError count ANode BError count BNode CError count CTrigger alert?
First runFail1Fail1Fail1NO
Second runGood0Fail2Fail2NO
Third run (Scenario I)Fail1Fail3Good0NO
Third run (Scenario II)Fail1Fail3Fail3YES

Note that error conditions are checked as each new data point comes in. In other words, as soon as a(ny) node reports data back to us, the node threshold (e.g., number of failing nodes) and consecutive failure settings are checked. The alarm alert is triggered or stays in active state as long as all criteria still hold true. If a node has a successful run, the error condition for that node is reset.