Agent Rules Help

Agent Rules Back to Settings
Manual Trigger

Run immediately from the rules table to investigate current incident context on demand.

Rule Trigger

Attach to anomaly or tag rule references and trigger when configured state transitions occur.

Rate Limited

Use per-rule cooldown and global assignment limits to avoid duplicate issue churn.

How Agent Rules Trigger
Rules can run manually or trigger from anomaly/tag rule states. Use rate limits to prevent repeated actions during noisy incidents.
Action Types
  • LLM analysis and suggested remediation
  • Optional DLP check before issue creation
  • Optional GitHub issue creation and Copilot handoff
Recommended Flow
  1. Create and test a rule in manual mode first.
  2. Attach one trigger reference (anomaly or tag rule).
  3. Set trigger state and conservative cooldown minutes.
  4. Enable analyze action, then add DLP/GitHub actions if needed.
  5. Review run history to tune frequency and action set.
Troubleshooting
  • If rules never run, verify trigger type/reference pairing.
  • If runs are skipped, check cooldown and global limit settings.
  • If GitHub issue creation fails, validate repository and token scope.
  • If analysis fails, verify primary and guard AI settings.
Agent flows depend on valid AI settings and, for GitHub actions, repository/token configuration.