Investigate detections
A detection uncovers unusual behavior on your network. After finding a detection, you might wonder how this behavior is directly affecting the devices and resources on your network. To find answers to your question, you can launch an investigation directly from a detection.
All detections include a general description of what contributed to the unusual behavior. For example, the following figure shows which client and methods are linked to the spike in server traffic volume.

To collect more specific information about a detection, you can continue your investigation by completing the following steps.
- For detections associated with L7 protocol metrics, click Activity map to see the peer device connections to the client or server associated with the detection.
- Click the device or application name to navigate to the protocol page.
- If you have a connected Explore appliance and detections associated with
metrics, click Records to view transaction-level
information filtered to the detection source.
Note: To view records, you must have stored records for the protocol associated with the detection. For more information, see Collecting and storing built-in records.
Open an activity map from a detection
When a single client or server is associated with unusual L7 protocol activity, such as a high number of HTTP errors or DNS request timeouts, an activity map link appears.
- Click any client in the map to access a menu that contains a Go to
Device... link. Click the link to open a protocol page with
client metrics, such as requests and responses.
- In the left pane below Step 1, click Add
Step and then click All Peers in the
drop-down list. The map updates to show you which downstream devices are
connected to the database clients, as shown in the following figure.
- Save and then share your activity map with other ExtraHop users.
For more information about activity maps, see Activity maps.
Navigate to a protocol page
If you want to further investigate anomalous metrics, you can navigate to a protocol page where you have access to additional charts, metrics, and tools.
Next steps
From a protocol page, you can then choose one of the following options to further investigate metric data:
Best practices for investigating detections
The Machine Learning Service provides you with high-quality, actionable data about detections—but does not replace decision-making or expertise about your network. The following best practices explain how to determine which detections are worth further investigation and when to take action.
- Change the time interval to see when detections occurred
- Learn if detections occurred before, after, or during a reported problem. For example, does
the time frame of the detection coincide with a reported issue, such as slow load times or
login times? You can also compare detections from the past month to the current date, which
gives you a sense of whether the occurrence or severity of detections is changing over
time.
For more information, see Filter the detections list.
- Monitor detection-related metrics in a dashboard chart
- You can create and edit a dashboard chart
from a detection. Click the sparkline, as shown in the following figure.
The detection source, metric, time interval, and drill-down details are preserved in the Metric Explorer so that you can quickly create a dashboard chart to monitor additional changes. - Create a detection alert
- You can configure an alert to receive email notifications when a detection occurs. Detection
alerts also help you quickly find detections for a specific device or application on the Alerts page page.
For more information, see Configure detection alert settings.
Thank you for your feedback. Can we contact you to ask follow up questions?