I've seen that in the past with Office365 Activity logs.

Any indication of what is happening in the job logs? Typically there's an HTTP timeout or network issue that might cause a failure. I'm curious what the logs say.

Getting a lot of 401s but then suddenly the jobs will start passing again. It just seems sporadic

I'd expect 429s not 401s. I wonder if the 401s are for files that's aren't finished writing yet

So maybe I should increase the polling interval?

It's at 10 minute intervals. Maybe I should try 20

Check the api docs also. It's strange if I remember correctly