Incident model feedback
Modify your incident model thresholds to get less noisy alerts.
Last updated
Modify your incident model thresholds to get less noisy alerts.
Last updated
To adjust the sensitivity of the incident alerting, you can give feedback to the model by heading into the Incident Details screen.
Decube's incident model is directly influenced by the confidence interval (CI). The confidence interval represents the system's required level of certainty before flagging an incident. By default, the CI is set to 0.90, meaning the system must be 90% confident that a detected anomaly is significant before triggering an alert.
You can click on the thumbs down icon to give the feedback rating. For an incident that has auto-thresholding enabled, you can then see a slider to adjust the sensitivity.
You can only adjust the sensitivity of the incident models for tests with automatic thresholds only, eg. Freshness, Volume, Field Health tests with automatic threshold selected.
To increase the sensitivity of the incident model (and therefore trigger more alerts), move the slider to the right. To reduce the sensitivity of incident model (and therefore reduce noisy alerting), move the slider to the left.
The maximum CI (on slider: -5) is 0.99 while the minimum CI (on slider: 5) is 0.80. There is an increase/decrease by 0.02 for each step in the slider. eg. slider position = 1, CI = 0.92.
Once you've moved the slider and submitted the feedback, subsequent monitor scans will follow the new setting for triggering alerts.
