Notifications
Send policy-violation alerts through email, Slack, or Microsoft Teams. Operators and Administrators can configure integrations; Viewers cannot.
What Triggers Notifications
Only policy violations trigger alerts. Build failures, discovery, policy edits, and integration health do not.
Violations are grouped by build or workload within each submitted batch. Later batches can produce additional alerts for the same run. Each matching integration delivers independently: email uses one Bcc message, while Slack and Teams post to each selected channel.
Available Channels
| Channel | Setup |
|---|---|
| Encrypted SMTP relay | |
| Slack | Slack app and workspace installation |
| Microsoft Teams | Azure Bot and Teams app installation |
How Notifications Are Delivered
The Sensor submits violations and Team Server queues delivery through its configured workers. Use BackgroundAsync; no notification scheduler job is needed. Delivery can start before a run finishes. ENDURA_BATCHER_AUTOFLUSH=true submits batches immediately and can increase alert volume.
An integration needs Active status, notifications enabled, a destination, and matching resource scope. Observed violations also require Include observed violations.
| Scope selection | Receives alerts for |
|---|---|
| All Pipelines | Current and future pipeline builds |
| Selected pipelines | Only those pipelines’ builds |
| All Deployments | Current and future deployment workloads |
| Selected deployments | Only those deployments’ workloads |
| No selection with the corresponding All option off | Neither that resource type nor its descendants |
Pipeline and deployment scopes are independent. Overlapping integrations both deliver.
BackgroundAsync holds queued work in memory. Shutdown can lose queued notifications; retries are bounded and failed deliveries have no durable replay queue. Check application logs and test delivery after recovery.
Who Can Configure Notifications
Operators and Administrators can view, create, edit, test, and delete integrations. See Access Control.
Next Steps
Follow the channel guide, test connectivity, then trigger an in-scope violation and confirm the alert links to the correct build or workload. Enable observed outcomes when testing an observe-mode policy.