Some of the most common workflows that you can perform through triggers include the following operations:
- Create an application container in which metrics are collected for specific devices. Application containers augment the device-based views that the ExtraHop system constructs by default.
- Create custom metrics and save them to the ExtraHop datastore. For example, user agent data generated by an HTTP request is not a metric built into the ExtraHop system. However, the ExtraHop Trigger API provides a user agent HTTP property, which enables you to write a trigger that collects user agent data as a custom metric.
- Generate records and write them to the ExtraHop Explore appliance for long-term storage and retrieval.
- Send data to syslog consumers, such as Splunk, or to third party databases, such as MongoDB or Kafka, through an open data stream.
- Perform universal payload analysis (UPA) to access and parse TCP and UDP payloads from unsupported protocols.
- Initiate packet captures to record individual flows based on user-specified criteria. Your ExtraHop system must be licensed for packet capture to access this feature.
To view all triggers, click the System Settings icon and then click Triggers. From the Triggers page, you can create a trigger or select the checkmark next to a trigger to edit the trigger configuration or modify the trigger script.
Plan a trigger
Writing a trigger to collect custom metrics is a powerful way to monitor your application and network performance. However, triggers consume system resources and can affect system performance, and a poorly-written trigger can cause unnecessary system load. Before you build a trigger, evaluate what you want your trigger to accomplish, identify which events and devices are needed to extract the data you need, and determine whether a solution already exists.
- Identify the specific information you need to collect, by asking the following types of
- When will my SSL certificates expire?
- Is my network getting connections on non-authorized ports?
- How many slow transactions is my network experiencing?
- What data do I want to send to Splunk through an open data stream?
- Review the Metric Catalog to determine whether a built-in metric already exists that extracts the data you need. Built-in metrics do not create additional load on the system.
- Identify which system events produce the data that you want to collect. For example, a trigger that monitors cloud application activity in your environment might run on HTTP responses and on the open and close of SSL connections. For a complete list of system events, see the ExtraHop Trigger API Reference.
- Familiarize yourself with the API methods and properties available in the ExtraHop Trigger API Reference. For example, before you get too far in planning your trigger, check the reference to make sure that the property you want to extract is available, or to find out what properties are collected in a default CIFS record.
- Determine how you want to visualize or store data collected by the trigger. For example, you can view metrics on a dashboard or by protocol, you can send records to the ExtraHop Explore appliance or other supported recordstore.
- Determine if a trigger already exists that meets your needs or might be easily modified; always start with a pre-existing trigger whenever possible. Search the following resources for an existing trigger:
If you determine that you need to build a new trigger, familiarize yourself with the following tasks that must be completed:
- Configure the trigger to provide details such as the trigger name and whether debugging is enabled. Most importantly, specify which system events the trigger will run on. For example, if you want your trigger to run each time an SSH connection is opened, you will specify SSH_OPEN as the trigger event.
- Write the trigger script, which specifies the instructions the trigger will carry out when a system event configured for the trigger occurs. The trigger script can provide instructions for a simple task such as creating a custom device count metric called "slow_rsp" or a more complex effort such as monitoring and collecting statistics about the cloud applications accessed in your environment.
After the trigger is complete and running, it is important to check that the trigger is performing as expected.
- View the debug log for expected output from debug statements in the trigger script. The log also displays any runtime errors and exceptions that you must fix.
- Monitor the performance cost by tracking the number of cycles consumed by the trigger.
- Check System Health charts for trigger exceptions, drops from the trigger queue, and unexpected activity.
- Check that the trigger script adheres to the Triggers Best Practices Guide.
Thank you for your feedback. Can we contact you to ask follow up questions?