Skip to main content
Use Tracee’s webhook output destination to send runtime security detections to Flashduty On-call. When the same detection rule fires again in the same container, the events merge into one alert. Tracee’s webhook sends detections only and no recovery notification, so alerts are closed by the channel’s auto-close or manually.

​
In Flashduty On-call

You can obtain an integration push URL in either of the following ways.

​
Use a dedicated integration

  1. In the Flashduty console, select Channel and open a channel
  2. Select Configuration → Integrations → Private integration, then click Add an integration
  3. Select Aqua Tracee, then click Save
  4. Open the generated integration card and copy the Push URL

​
Use a shared integration

  1. In the Flashduty console, select Integration Center → Alert Events
  2. Select Aqua Tracee and enter an integration name
  3. Configure the default route and select a channel; after creation, add more rules under Route if needed
  4. Click Save and copy the generated Push URL

​
Configure Tracee

1

Add a webhook output destination

Tracee defines output destinations with the --output flag or the output: section of its configuration file. Put the Flashduty push URL in url and keep the default json format. The webhook POSTs every event to that URL as application/json, with no signature and no extra headers needed.Command line:
Configuration file:
2

Send detections only

Flashduty creates alerts only for detection events, which carry a threat field. Events without threat (for example sched_process_exec) are acknowledged and dropped. To avoid sending every event to Flashduty, use a stream so that only detection events reach this destination:
List the detection events you want to alert on under events (events produced by a detector or signature, such as anti_debugging).
Use the default json format. With a custom gotemplate= template, Flashduty still parses the default JSON field names, so missing fields lead to missing labels or alerts that merge together.
3

Turn on auto-close

Tracee’s detections have no recovery notification: after a threat is handled, Flashduty receives nothing.Turn on auto-close in the channel that receives these alerts, with a suggested duration of 24 hours. When the same rule fires again in the same container within the grouping window, the same alert is updated; after the alert is closed, the next detection opens a new one.
4

Verify

Tracee has no “send test notification” feature. Trigger a real detection in Tracee, or send a sample detection to the push URL:
Confirm that an active alert appears in Flashduty with threat.name as the title. The sample opens a real alert, so close it manually after verifying.

​
Alert Key

Tracee events have no unique ID. Flashduty builds the Alert Key from the rule identity plus where the event ran (the MD5 of "tracee", the rule identity and the location, separated by NUL):
  • Rule identity: the first non-empty of threat.properties.signatureID, threat.properties.id and the event name. Tracee documents threat.properties.signatureID as the signature ID of a detection; events from newer detectors do not carry it, so the event name stands in
  • Location: workload.container.id for an event inside a container; otherwise workload.process.host_pid, then workload.process.pid
So repeated hits of the same rule in the same container merge into one alert, while different rules or different containers stay separate alerts. Changes to the event time, process name, event data, severity and threat.name do not change the Alert Key.
Events outside a container are told apart by process ID only. Tracee events carry no hostname, so the same rule on two hosts with the same process ID merges into one alert; in that case use a separate integration per host.

​
Severity

In the default JSON, threat.severity is a number (INFO=0 to CRITICAL=4):

​
Labels

Flashduty sets these labels when they have a value: rule_id, signature_id, event_name, category, threat_severity, mitre_tactic, mitre_technique_id, mitre_technique, container_id, container_name, image, pod, namespace, process, executable, pid, host_pid, policies, detected_from. The event’s data array (file paths, command arguments and so on) is never copied into labels.

​
Troubleshooting

  • Tracee logs Error sending webhook, http status: confirm the URL is complete and includes integration_key
  • No alerts arrive: confirm the event carries a threat field; ordinary events create no alert, and the stream’s events must include the detection event
  • Alerts do not recover: Tracee sends no recovery notification; turn on the channel’s auto-close or close alerts manually
  • Alerts from different hosts merge: events outside a container are told apart by process ID only; see the note under Alert Key
For field details see the Tracee outputs documentation and Event structure.