[04:11:53] where do errors like "Error: 403, Too Many Requests (ce20de6)" come from? [04:31:19] mostly I'm wondering if I should be treating it like 429 and just keep retrying or something else [04:31:26] and also how to avoid hitting it... [16:52:58] legoktm: retrying won't work, what were you doing when you hit it? [17:27:39] Raine: just running automated tests for my bots/Toolforge tools locally that use a few different APIs [17:29:58] also seeing it in GitLab CI, e.g. https://gitlab.wikimedia.org/repos/mwbot-rs/mwbot/-/jobs/705056 [17:37:20] legoktm: are your tests running in a browser? or the rust image? [17:39:20] not a browser no, just Rust code [17:39:53] cool, thanks [17:40:26] should be okay now [17:43:15] Raine: still getting them, e.g.: [17:43:20] thread 'media::test::test_commons_url' (62066) panicked at src/media.rs:132:14: [17:43:20] called `Result::unwrap()` on an `Err` value: HTTP status client error (403 Too Many Requests (ce20de6)) for url (https://upload.wikimedia.org/wikipedia/commons/thumb/f/f1/Pillars_of_Creation_%28NIRCam_Image%29.jpg/1247px-Pillars_of_Creation_%28NIRCam_Image%29.jpg) [17:44:04] right, my change might take a few minutes to propagate, sorry about that [17:44:36] Raine: note that's upload, I think the rule you touched is scoped to text only [17:45:10] oh, you're right taavi [17:47:50] thanks '^^ also applied to upload now [17:48:34] (also, hi legoktm!!) [17:50:17] o/ hiii [17:51:46] retriggering the mwbot-rs CI job now, let's see if it passes :fingers_crossed: [17:52:18] I just started my winter break from school so now I'm just trying to get everything in a healthy state before I disappear again [17:53:26] happy disappearance :-) [17:55:16] yay https://gitlab.wikimedia.org/repos/mwbot-rs/mwbot/-/jobs/705447 all passed [17:55:18] thank you Raine!! [17:55:31] \o/ [17:55:53] thank _you_ for poking at it enough to find the problem :D [17:57:49] :) [17:58:17] legoktm: oh, and as for how to avoid it: you got caught in someone else being mean, nothing you can do about that, it really helps when you report it to us (either here, feel free to ping me if you don't get a response, or send an email to noc@) [18:02:51] got it I think I was just mostly confused by "Too Many Requests" usually being a 429 rate limit thing and not an abuse thing [18:04:37] yeah, in this case the someone was being extra mean :D [18:06:31] we're working on improving our systems so legitimate users don't get caught in things when we need to take extra measures, but there is a _lot_ of scrapers out there now, so things won't always be perfect, sorry about that [18:08:27] but yeah, you're correct that with 429s retrying should work, but if you get a 403, it generally won't go away on its own [18:12:02] *nods* [18:12:38] and no worries, I get it! :) [18:13:55] :-)