Troubleshoot integrations
Separate connection, authentication, payload, delivery, and business-result failures.
Diagnose in order
- Confirm the business, integration, provider account, and active state.
- Run the available connection test. Record its exact error rather than repeatedly changing unrelated settings.
- For authentication failures check expiry, key state, scopes, and provider consent. Regenerate or rotate a credential only after identifying all consumers.
- For event failures compare direction, event name, and payload fields with the configured consumer. Inspect delivery history and workflow conditions.
- For wrong product data compare external ID, SKU, source, and last sync time.
- For a successful request with no result inspect downstream runs, tool output, and the provider record.
- Before retrying, check whether a previous attempt already sent a notification or changed data.
Escalate with useful evidence
Share the time and time zone, integration ID, event or execution ID, sanitized error, expected result, and actual result through support. Remove full API keys, authorization headers, mailbox passwords, and unnecessary customer information.
Expected result
You can identify which stage failed and reproduce it with controlled data. Use channel troubleshooting for inbox and delivery issues.
Diagnose missing messages and disconnected channels
Trace account selection, provider connection, inbox state, and delivery errors.
Monitor and recover product synchronization
Check source ownership, last sync, integration health, and failed updates.
Configure inbound and outbound webhooks
Receive source events, notify external systems, and inspect delivery results.