PagerDuty

Gate AI-driven PagerDuty incident response actions — automated runbooks, service restarts, escalation reassignments — behind a human approval step before they execute.

Prerequisites

PagerDuty account with Event Intelligence or AIOps add-on.
A backend endpoint or serverless function to act as middleware.
Cheqpoint Connection Key.

Steps

In PagerDuty, go to Integrations → Generic Webhooks (v3) and create a webhook that fires when an AI-driven automated action is about to execute.
Route the webhook to a middleware endpoint (e.g., a serverless function) that you control.
In your middleware, POST the incident details to https://cheqpoint.co/api/webhooks/inbound with the x-api-key header.
Block the automated PagerDuty action until Cheqpoint returns a decision.
If approved, trigger the runbook or automated response via the PagerDuty API.
If rejected, suppress the automated action and create a manual on-call task for the engineer.

Sample request payload

json

{
  "action": "run_auto_remediation_runbook",
  "summary": "AI requesting automated DB failover for incident INC-4421",
  "details": {
    "incident_id": "INC-4421",
    "service": "payments-db",
    "runbook": "db_failover_primary_to_replica",
    "severity": "P1",
    "affected_regions": ["us-east-1"]
  },
  "justification": "Primary DB CPU at 98% for 8 minutes. Automated diagnosis suggests failover."
}

Sample Cheqpoint response

json

{
  "status": "approved",
  "modifiedDetails": null,
  "decisionNote": "On-call confirmed. Run failover. Alert #payments-team Slack channel."
}

Async behaviour

json

// For time-critical incidents, use callbackUrl to avoid blocking the middleware:
//
//   a) Submit the approval request with callbackUrl
//   b) Return a 200 to PagerDuty immediately (suppress the automated action for now)
//   c) When the reviewer decides, Cheqpoint POSTs the decision to your callbackUrl
//   d) Your callback handler then triggers or suppresses the runbook via PagerDuty API

{
  "action": "run_auto_remediation_runbook",
  "summary": "AI requesting automated DB failover for incident INC-4421",
  "details": { "incident_id": "INC-4421", "runbook": "db_failover_primary_to_replica" },
  "callbackUrl": "https://your-backend.example.com/cheqpoint/pagerduty-callback"
}

Notes

For P1 incidents, configure short SLA targets in Cheqpoint (e.g. 5 minutes) so the approval request auto-escalates if no reviewer responds in time. Set the on-expiry action to approve or reject based on your incident response policy.

Tips

Start by routing only high-risk or high-value actions to minimize friction while maintaining oversight.

Get your Connection Key at cheqpoint.co/signup.