The hackathon is in Turkey with some amount of restriction on the ISP/network. Some things won't load...
Description
Related Objects
- Mentioned Here
- T353122: GitLab login 422: email has been taken
Event Timeline
Cannot reproduce. I'm on the Wikimedia wifi at the Hackathon and I just logged into our GitLab without a problem, and 2FA also worked without a problem.
FWIW, I'm not seeing much of a spike in 4xx responses per grafana: https://grafana.wikimedia.org/d/R_1IvBZnz/gitlab-omnibus-overview?orgId=1&refresh=1m&viewPanel=38
Wonder if there are more specific auth logs some place...
First thoughts by some might be related to throttling (fyi, for those with access, this has the details where we exempted the Hackathon network T382309).
But this isn't a 429, it's a 422. (.."indicates that the server understood the content type of the request content, and the syntax of the request content was correct, but it was unable to process the contained instructions..")
Searching Phabricator for previous "422" along "OIDC" found this one T353122.
There this was for "email already taken".
I had an issue earlier, so did another person…
I’m not sure if either was after the wifi reset
I can confirm though there are multiple hits in the gitlab access_log that match a 422 and an IP from the Hackathon network.
Here are the timestamps. There is block between 09:31 and 09:43, another 10:04 to 10:24, 2 lone events at 15:34 and then it stopped. Would this correlate with local network events?
@cee: {"timestamp":"2025-05-02T09:31:31+00:00"
@cee: {"timestamp":"2025-05-02T09:32:12+00:00"
@cee: {"timestamp":"2025-05-02T09:33:58+00:00"
@cee: {"timestamp":"2025-05-02T09:34:36+00:00"
@cee: {"timestamp":"2025-05-02T09:36:14+00:00"
@cee: {"timestamp":"2025-05-02T09:36:37+00:00"
@cee: {"timestamp":"2025-05-02T09:37:15+00:00"
@cee: {"timestamp":"2025-05-02T09:43:16+00:00"
@cee: {"timestamp":"2025-05-02T10:04:04+00:00"
@cee: {"timestamp":"2025-05-02T10:04:16+00:00"
@cee: {"timestamp":"2025-05-02T10:04:25+00:00"
@cee: {"timestamp":"2025-05-02T10:04:55+00:00"
@cee: {"timestamp":"2025-05-02T10:05:07+00:00"
@cee: {"timestamp":"2025-05-02T10:05:12+00:00"
@cee: {"timestamp":"2025-05-02T10:06:56+00:00"
@cee: {"timestamp":"2025-05-02T10:07:04+00:00"
@cee: {"timestamp":"2025-05-02T10:08:21+00:00"
@cee: {"timestamp":"2025-05-02T10:08:51+00:00"
@cee: {"timestamp":"2025-05-02T10:09:02+00:00"
@cee: {"timestamp":"2025-05-02T10:09:12+00:00"
@cee: {"timestamp":"2025-05-02T10:09:32+00:00"
@cee: {"timestamp":"2025-05-02T10:09:37+00:00"
@cee: {"timestamp":"2025-05-02T10:11:20+00:00"
@cee: {"timestamp":"2025-05-02T10:12:03+00:00"
@cee: {"timestamp":"2025-05-02T10:12:07+00:00"
@cee: {"timestamp":"2025-05-02T10:12:27+00:00"
@cee: {"timestamp":"2025-05-02T10:14:29+00:00"
@cee: {"timestamp":"2025-05-02T10:14:47+00:00"
@cee: {"timestamp":"2025-05-02T10:15:32+00:00"
@cee: {"timestamp":"2025-05-02T10:15:45+00:00"
@cee: {"timestamp":"2025-05-02T10:15:55+00:00"
@cee: {"timestamp":"2025-05-02T10:16:24+00:00"
@cee: {"timestamp":"2025-05-02T10:17:09+00:00"
@cee: {"timestamp":"2025-05-02T10:19:32+00:00"
@cee: {"timestamp":"2025-05-02T10:20:21+00:00"
@cee: {"timestamp":"2025-05-02T10:20:25+00:00"
@cee: {"timestamp":"2025-05-02T10:20:43+00:00"
@cee: {"timestamp":"2025-05-02T10:21:52+00:00"
@cee: {"timestamp":"2025-05-02T10:23:04+00:00"
@cee: {"timestamp":"2025-05-02T10:23:15+00:00"
@cee: {"timestamp":"2025-05-02T10:23:18+00:00"
@cee: {"timestamp":"2025-05-02T10:24:33+00:00"
@cee: {"timestamp":"2025-05-02T15:34:26+00:00"
@cee: {"timestamp":"2025-05-02T15:34:51+00:00"The url.path is almost always /users/auth/openid_connect/callback?... with the exception of the last 2 which were for import/github.json.
The referer is mostly just gitlab.wikimedia.org, a few are idp.wikimedia.org and the last 2 are gitlab.wikimedia.org/import/github/status.
Response body.bytes is always 49847 with the exception of the last 2 206.
The user agent is all either Mozilla/5.0 (X11; Linux x86_64; rv:132.0) Gecko/20100101 Firefox/132.0 or Mozilla/5.0 (X11; Linux x86_64; rv:135.0) Gecko/20100101 Firefox/135.0. Not a single other one for all of those.
The only exception again the last 2 hits where it was Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36.
It has also been confirmed that our CAS (idp) has mostly been idling.
So it does seem more likely related to local network.
