Skip to main content
Start from the observed symptom, confirm the failure stage, and apply the matching remedy. A consistent stage-first diagnosis avoids chasing the wrong component.

Goal

Diagnose webhook delivery, SQL collection, authentication, processing, gate, and feedback failures by symptom.

Prerequisites

Start from the observed symptom and confirm which integration stage failed.

Symptom reference

Stage-by-stage diagnosis

Webhook delivery

  • Confirm the endpoint is reachable and accepts only the expected sources.
  • Verify the shared secret or signing key matches on both sides.
  • Check DNS, proxy, firewall, and TLS certificate validity.
  • Inspect retries and delivery logs on the source platform.

SQL collection

  • Confirm the diff includes only new and modified SQL files.
  • Verify path and extension filters, and that removed files are ignored.
  • Check that the workspace dialect matches the SQL being reviewed.

Authentication

  • Confirm the credential is valid, unexpired, and scoped to the target.
  • Verify that the endpoint allowlist includes the calling platform.
  • Rotate the credential and re-run if the error persists.

Processing

  • Check that the workspace has the DDL or connection needed by the policy.
  • Confirm the review policy is published and applicable.
  • Look for parse errors caused by a dialect or script mismatch.

Gate

  • Confirm the applied policy version and blocking criteria are correct.
  • Review the failure policy for timeout and unavailability cases.
  • Inspect manual exceptions for scope, approver, and expiration.

Feedback

  • Verify the status or comment permissions on the source platform.
  • Confirm the result targets the current revision.
  • Check that duplicate delivery is not creating conflicting feedback.

Expected Result

The failure is traced to a specific stage and the matching remedy from the stage-by-stage diagnosis is applied.