M365 SharePoint backups repeatedly fail with HTTP 429 when another backup product protects the same tenant

M365 SharePoint backups repeatedly fail with HTTP 429 when another backup product protects the same tenant

Machine-distilled from a resolved support ticket on 2026-09-28. Source ref: dc75de9b91d4. Verify before relying on it.

Applies to: HYCU R-Cloud protecting Microsoft 365 (SharePoint Online and Exchange Online); version not specific

Symptom: Large numbers of backup failure notifications when protecting Microsoft 365. Exchange sources are protected, but some SharePoint sites (typically very large ones) stay unprotected because their initial backups keep failing or completing with errors. Task reports show only HTTP 429 (throttling) errors, and event-based notifications produce many alerts for the same underlying failure.

Cause: Microsoft applies Graph API throttling at the tenant level, so all applications calling Graph in the tenant share the same limits. When a second (legacy) backup product is still running against the same tenant, combined API consumption exceeds the limit, and Microsoft returns HTTP 429. Very large SharePoint sites (several terabytes) need many API calls for their first full backup, which makes throttling far more likely. Each API retry raises a separate event, which inflates the number of notifications.

Resolution: 1. Review the task reports for the failed SharePoint backups and confirm that the errors are HTTP 429 throttling rather than another failure type. 2. Check whether any other backup or migration tool is using the Microsoft Graph API against the same tenant. If one is, pause or disable it so HYCU has enough API capacity. 3. Allow the initial full backups of large sites to finish. They may take several days and may still log throttling errors before they complete. Incremental backups should then succeed. 4. For monitoring, filter the dashboard for unprotected entities and review task-level reports, rather than relying on event-level notifications that raise an alert for every API retry. Only investigate entities that fail on consecutive attempts, because single intermittent failures are expected. 5. If throttling continues after the other product has been paused, escalate to HYCU with the sizes of the affected sites.