Token refresh flow causing 504 gateway timeouts and 500 internal server errors

[mgoodfellow]

Hi,

We have had a spike of errors from token refreshes returning GatewayTimeout 504 or Internal Server Error 500s. No other endpoints appear to be affected.

This has the unfortunate side effect of logging out large numbers of users and forcing them to reauthenticate. We cannot know if the token has been consumed or not when we face these kinds of errors and have to fail safe.

Issue started at 14:37 UTC, and seems to have calmed down around 14:47 UTC.

Thanks!

[mgoodfellow]

No more occurrences of this issue, so I guess it was a transient issue / bad deploy!

Closing this as it isn't ongoing.

Cheers!

[mgoodfellow]

This issue has started against 20:27 UTC and continued until 20:34 UTC.

[mgoodfellow]

As before, we are only seeing failures on the token refresh endpoint - but it is causing a bad user experience for us and our users. Was a change made to authentication service today?

[dasha-kobzeva]

Hey @mgoodfellow,
we had some latency spikes to our auth services, team is investigating the cause of it.