The Alerts page is where you can monitor all of the notifications you’ve received for anomaly detection workflows.
These are triggered when our algorithm detects a value far outside what we’d usually expect. You can learn more about how anomaly detection works here.
Anomaly detection workflows are created in the Workflows page. Clicking ‘Create Workflow’ from the top right-hand corner within the alerts page will take you to the Workflows page.
Alerts Menu
This view shows you all of the alerts that have been sent whenever an anomaly has been detected in your data. From the menu, you can see details about when the alert was sent, the workflow that caused the Alert to be sent, and a short description of what the Alert was measuring.
From the drop-down menu in the top right-hand corner, you can adjust the period to see Alerts sent across a selected time frame.
If there’s a new Alert for you to look at, you’ll see a small red dot next to the Alerts menu, so you know there’s something to check up on.
Track the status of Alerts
Alerts help your team stay transparent over what’s been attended to, by allowing you to mark individual alerts with either an ‘Open’ or ‘Reviewed’ tag.
When Alerts first appear, they’ll be automatically assigned the ‘Open’ tag. Clicking on this will give you the option to mark it as ‘Reviewed’. These changes will appear for everyone in your team, so you can stay up-to-date with who’s attending to different issues.
You can also sort all of your alerts by those that are ‘Open’ or ‘Reviewed’ by selecting from the filters at the top of the table.
Drill down into individual Alerts
If you Click on an individual alert you’ve received, you’ll be brought to a new menu where you can look into more details about each anomaly.
Alerts show you the anomaly against a graph of your data from the last two weeks, with the anomalous result highlighted with a red line. You can hover over individual data points to see their value too.
Below the Alert graph, Insight Assistant lists key highlights from the feedback data that this Alert has triggered within. You can also click into 'Feedback' to see the raw customer comments that are causing this Alert to send.
From the right-hand sidebar, you can access a range of details about each Alert.
Occurred on: The date the alert was spotted
Anomaly: The value of the anomaly that was outside of the expected range
Expected Range: The range of values we’d usually expect for this filter set
Theme: If you’ve selected a theme to track in the filters for this Alert it’ll appear here
Data Source: If you’ve selected a specific data source in filters for this Alert, it’ll appear here
The Alert Workflow section gives you details about the workflow that triggered the alert to send. You can click on the cog icon to jump into workflows, where you can view and edit the workflow itself.
Name: The name of the alert created
Metric: The metric you’re judging the anomaly off, if you’ve selected one
Frequency: The window you’ve set for the Alert to be judged on This can be Daily, Weekly, or Monthly. Learn more about Frequency here.
Sensitivity: The threshold required for the alert to send. It can be high, medium, or low. Learn more about Sensitivity here.
Notified: Lists the email addresses, slack profiles & channels which the alert was sent to.
Sending Alerts to Jira
Via the top left-hand corner of the Alert viewer, you can click to instantly turn any Alert into a Jira ticket. Clicking on this will automatically generate a ticket in Jira, with the following details about the Alert included:
A summary of the feedback causing the Alert to send & associated metadata
A link back to the original Alert in Chattermill