Our service is 100% on token-based authentication with APNs. Each pod generates a new token to use every 30 mins. We have hundreds of pods.
We implemented this way because in Refresh your token regularly section it says
Refresh your token no more than once every 20 minutes and no less than once every 60 minutes.
APNs report an error if you use a new token more than once every 20 minutes on the same connection.
However, we saw TooManyProviderTokensUpdates sporadically. What is even stranger is, the trigger of the error does not correlate to the number of pods as we scaled up or down, nor does it correlate to the QPS. To me, it looks like It triggers randomly. Can someone from Apple shed some light on this?