503 errors with OneDrive usually point to a few common issues:
1. Microsoft API rate limiting
OneDrive/Microsoft Graph API has rate limits. If you’re making many requests in succession, you might be hitting throttling. The “Service unavailable” message is Microsoft’s way of saying “slow down.”
Quick fixes:
Add a Wait node (1-2 seconds) between OneDrive operations
If processing multiple files, batch them with delays
Check if other workflows are hitting the same OneDrive account simultaneously
2. Large file downloads
If you’re downloading large files, the default timeout might not be enough. In the HTTP Request node settings (or OneDrive node if using that), you can increase the timeout value.
3. Microsoft service issues
Sometimes it’s just Microsoft having a bad day. Check Microsoft 365 Service Health to see if there are ongoing issues.
4. Node retry settings
As the error message suggests, you can enable automatic retries in the node settings:
Go to node Settings → Error Handling
Enable “Retry on Fail”
Set retry count (e.g., 3) and wait between retries (e.g., 5000ms)
Quick diagnostic:
Is this happening with all files or specific ones?
Did anything change recently (file sizes, frequency of runs)?
Does it work if you run the workflow manually with a single file?