We have been seeing significantly more 401 errors over the past 48 hours. Calls to account/verify_credentials succeed, but calls to statuses/mentions or statuses/direct_messages fail are failing with authentication errors.
There’s a definite possibility of a rise in 401s right now; current working theory is that these are 500-series errors incorrectly being raised as 401s and we’re working to resolve that.
Is there an open bug on this? We are having this issue across several of our client accounts. I can provide a set of reproducible steps if that would help you all to diagnose and fix the underlying issue.
I’ve opened a bug for this that you can track. If there’s something deterministic you’ve found to cause the issue, please do let us know – we believe we’ve identified the problem and are working to fix it, but determinism isn’t a component to it from what I can tell.
Here’s the issue:
Can we get an update on this bug?
I think I am currently experiencing this issue. All my GET requests (with 2 different app) to home_timeline, mentions result to a 401 response while POST requests like update works fine (with the same credentials)
Same weird behaviour with realtime API : userstream/user --> 401 but stream/filter --> 200
With another account (so other credentials), but on the same computer (& same app) all work fine.
Finally, looking at my users, today is the worst day with a lot of 401 responses on userstream/user ( 5 times more than yesterday ).
Any update on this?
I don’t yet have a more granular ETA as to when this issue will be fixed. We’re still working to resolve it. I’ll update the thread and the ticket when I have more information.
If you know, please tell me the conditions in which this problem occurs.
Because if this problem is not solved for several more weeks, I have to avoid this problem by a certain method.
We’ve made some adjustments and think you should begin seeing these spurious 401s diminishing now. Let us know if you continue seeing the problems – we’ll hold out a bit before closing the ticket.
This issue is as yet not been resolved.
There are reports from users of my clients.
Until the fix applied to all users take a long time?
Same here … Still getting 401s for many users with valid authorization data (as they’ve just registered today).
Though, congrats on the architecture updates!
Still getting 401s for many users with valid authorization data too.
Yesterday I had the problem with my own account for about 4 hours. Then without reauthorizing, it works, just ununderstandable.
I don’t know if it can help but here is the number of 401 event received from usertstream API (few of them may be due to bad token) : http://awesomescreenshot.com/0a46vlx21
Thanks for the reports, we’re continuing to investigate.
I’m also getting 401 errors seemingly randomly mixed with 200 codes for identical authorization data, in calls to user_timeline.
A weird temporary fix for user is to Connect to www.twitter.com, Go on Settings, No change but just saved the settings --> At that point, the 401 errors disappear until the next occurrence.
Thanks for the reports everyone. We’re still actively investigating these. All these details really help and we appreciate it.
We are having these 401 issues also. Most of today.
Any time frame for a fix?
Yes, I’m getting this problem (this account), with streaming. It started 2012-06-10 13:47:04 UTC, and all reconnects since have failed. I reset keys, but it did not help. If the 401 should be a 5xx error, I’m wondering what it could be, as it was working fine for the three days previous to the above, and nothing has been touched.
Incidentally I also tried clicking Recreate my access token, for this application, but the access token is not showing as changed (I’ve even logged out and back in, after 20+ minutes). Could that be related?