Use this checklist after configuring either the direct Power Automate recipe or the Azure-verified recipe. Test in a separate Outlook calendar with disposable Tempo plans. Keep unrelated appointments and existing production connections unchanged.
Before the first test
-
Finish your recipe's authentication setup and choose the intended Outlook connection, calendar and timezone.
-
Use Daily events with start and end times and include created, updated and deleted planning events. Keep the source key and complete occurrence ID stable.
-
Keep automatic delivery disabled during preparation. Enable only the isolated test destination and required listeners when you are ready for the controlled lifecycle test.
Check the complete lifecycle
-
Create one disposable plan. Inspect the matching Power Automate run and actual Outlook appointment: calendar, date, start/end times, timezone and description must match.
-
Send the same occurrence again. There must still be exactly one matching appointment. Change its time and description in Tempo, then send the change: the same appointment should update.
-
For a multi-day plan, remove the first or last day and re-add a day. Check the full occurrence IDs, removed appointments and retained dates.
-
Delete the test plan. Confirm that only its managed appointments disappear. For recurring plans, also check the expected date range and exclusions.
-
If using preview or an initial sync, follow Preview planning sync. Confirm discovery coverage, inspect skipped/failed items, and compare the calendar after submission.
A successful import proves only that Power Automate accepted the package. A webhook acknowledgement alone does not prove the final calendar state. Sequential repeat checks do not establish atomic duplicate prevention for simultaneous first deliveries; validate receiver capacity and concurrency before a large rollout.
Check authentication separately
Use the authentication checks in the direct recipe or Azure recipe you configured. Direct URL-only and URL-plus-secret configurations differ from Azure HMAC verification and managed-identity access. Do not change a trigger to Anyone to troubleshoot an Azure identity failure.
Troubleshoot without replaying blindly
|
Symptom |
First check |
|---|---|
|
No flow run or 401/403 |
Return to your recipe's URL and authentication steps. Check the intended enabled flow and, for Azure, the verifier configuration and identity allowlist. |
|
Wrong dates, calendar or duplicate matches |
Check calendar/timezone configuration and complete occurrence matching. Preserve unrelated appointments and agree a scoped correction before retrying. |
|
Throttling, timeout or uncertain result |
Wait for the exact flow run to finish, then inspect Outlook. A response timeout can happen after a calendar write; do not resubmit the whole sync as the first remedy. |
For status meanings, failed-item retry and reconciliation procedures, use Delivery status and recovery.
Finish the test and retain useful evidence
Remove only the disposable plans and managed appointments you created. Return the isolated flow and destination to the agreed disabled state. Retain package versions/hashes, sanitized run IDs, expected/observed counts and the test outcome. Do not include signing secrets, full signed URLs, tokens or private calendar bodies in screenshots or support messages.
Return to the Recipes index to choose or review your connection setup.