As you start sending your metrics to Anodot, our algorithms will automatically start learning them to determine their normal range (aka ‘baseline’). Any deviation from this normal pattern will be detected in Anodot’s system as an Anomaly.
Naturally, not all anomalies are expected to be equally informative. In order to be notified only on relevant anomalies, you will first need to create an Alert. The alert will trigger once a set of conditions has been satisfied, and you will be notified on it via a dedicated channel (email, Slack etc).
A repository of all triggered alerts is available to you in the Alert Console. Here, you will also find two feedback buttons: Good Catch and Not Interesting.
By logging feedback on triggered alerts, you are helping us to provide you with a better experience and value, because:
- All feedback is fed into our Recommendations engine. Once enough feedback has been provided for a given alert, automatic recommendations will appear and suggest how to tweak the alert settings to reduce its false-positive rate (see more details here).
- Your feedbacks are routinely reviewed by our Customer Success team, who may suggest advanced fine-tuning and best practices for even better performance and efficiency.
To provide our team with a better understanding of why was the alert ‘not interesting’, it is highly recommended to indicate one of the following reasons:
- No Business Impact - A real anomaly, yet not relevant from a business perspective
- Incorrect Data - The data in Anodot is not valid / aligned with your database records
- Not an Anomaly - This is the normal behavior of the data, Anodot should have learned it
And, while this is optional, we strongly encourage you to provide us with any other relevant information:
Wait! there’s more!
Here are a few other ways to provide feedback on Anodot alerts:
- While investigating the incident in the 'anomalies' page:
- From within the email message:
- From within the Slack notification: