Create anomaly detection rules
Create anomaly detection rules to detect outliers in your collected data indicative of user or system behaviors that deviate from the norm so that you can further investigate and take preventive actions on potential problems. An anomaly detection rule is defined as a custom saved search scheduled to run on a regular basis. Once created, the recurring job shows in the Anomaly Detection Jobs panel in the Anomaly Detection Overview dashboard.
- In the Web UI, choose Search > Search.
- Enter a search to use as anomaly detection rule and run the search. For example,
sourcetype=aws:cloudtrail | timechart count span=30m
.
Note: A valid anomaly detection rule search must contain the... | timechart count ...
search fragment that produces time-series data. - Go to the Visualization tab and choose the Anomaly Detection Visualization chart.
- Click Schedule Job.
- In the Anomaly detection job settings window, schedule the anomaly detection job and enter information such as priority and train period.
By default, saved searches run at 5 minutes past the hour on an hourly basis. Use the same frequency for alerts and make sure alerts are triggered after the search job is complete, taking into account the search execution time. By default, the alert is triggered at 15 minutes past every hour, 10 minutes after the search job is scheduled to run. - Click Save.
Filter dashboards by tags in the Splunk App for AWS | Topology dashboard reference for the Splunk App for AWS |
This documentation applies to the following versions of Splunk® App for AWS (Legacy): 5.1.0, 5.1.1, 5.1.2, 5.1.3, 5.2.0, 6.0.0, 6.0.1, 6.0.2, 6.0.3
Feedback submitted, thanks!