[03:42:15] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2378215 (Antigng_) Labs replicas can't do that job, as revision tables are removed on such databases. Dumps are not updated such often. [07:18:54] DBA, Operations, Wikidata, Performance: EntityUsageTable::getUsedEntityIdStrings query on wbc_entity_usage table is sometimes fast, sometimes slow - https://phabricator.wikimedia.org/T116404#2378551 (adrianheine) [09:51:53] jynus: Thanks for the new s5 slaves… they already noticeably outperform the old ones :) [09:52:15] oh, how did you see that? [09:52:38] I do not have them yet at 100% target load [09:53:03] but it is the intention to get rid of all 64G servers [09:54:20] jynus: The (wikidata specific) dumps (that only load latest revisions, so same queries as via UI/API) are way faster if they happen to pick one of the new slaves [09:55:01] the truth is that s5 was a bit low on resources since one of the largest servers got fried [09:55:28] I have one issue with those dumps [09:55:43] and it is that they are long-running, but do not check cofiguration often [09:55:53] Yeah… I also don't like that [09:55:54] that leads to issues on depool [09:56:02] Not only that, but thaht as well [09:56:16] MediaWiki doesn't implement any kind of configuration re-checking [09:56:20] config should be realoaded at least every few minutes/hours [09:56:24] and that would also be hard because we do PHP "magic" [09:56:33] yes, it is not a wikidata specific thing [09:56:59] but in other cases, long running goes just to vslow/dump, so it hides the issue [09:57:08] true [09:57:15] I do not know if it is worth revisiting it on mediawiki [09:57:36] I think I will implement connection pooling/proxying earlier [09:57:36] We could potentially do the dumps in very small slices, thus a new process every few minutes [09:57:48] But connection proxying would be nicer [09:58:23] we had a related issue with video scalers [09:58:56] there was not a way to "disconnect" and then "reconnect" during encoding [09:59:05] and they kept master connections for a long time [10:01:12] regarding latency, I also intend to add to graphite soon the average query latency for all servers [10:01:13] our dumpers also sometimes open master connections [10:01:15] not sure why [10:01:37] But I have a suspicion… guess that could be fixed [10:01:39] that will help with the weights [13:32:31] DBA: Tendril DB graphs say 24hr but at 12hr - https://phabricator.wikimedia.org/T137654#2379525 (jcrespo) p:Triage>Lowest [13:34:36] DBA, Labs, Labs-Infrastructure: Add shared edit rights between eranbot and copypatrol tools - https://phabricator.wikimedia.org/T137691#2379529 (jcrespo) p:Triage>Normal a:jcrespo [13:39:43] DBA, Labs, Labs-Infrastructure: Add shared edit rights between eranbot and copypatrol tools - https://phabricator.wikimedia.org/T137691#2379537 (jcrespo) I am sorry, with this petition it is not completely clear to me: 1) which is the name of the database that you need to edit? 2) which of the users... [13:40:13] DBA, Labs, Labs-Infrastructure: Add shared edit rights between eranbot and copypatrol tools - https://phabricator.wikimedia.org/T137691#2379538 (jcrespo) [13:46:47] DBA: Tendril DB graphs say 24hr but at 12hr - https://phabricator.wikimedia.org/T137654#2374466 (jcrespo) To clarify, tendril's graphing capabilities are to be deprecated in favor of a more integrated diamond + graphite/prometheus + grafana solution, that is why this has very little priority. Related: T12675... [13:48:37] DBA: Tendril DB graphs say 24hr but at 12hr - https://phabricator.wikimedia.org/T137654#2379577 (jcrespo) Also, "5 minutes" and "1 hours" as units is a really bad idea in general, and makes impossible to compare those graphs. The unit should be seconds or minutes in both cases. [15:36:42] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2379894 (Antigng_) If you don't give me a good reason why cp1008.wikimedia.org:3128 / index.php?action=raw shouldn't be used, I will start some of my jobs that don... [15:44:55] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2379907 (Joe) @Antigng_ just to understand, what is your bot doing? If dumps are not refreshed fast enough for you, maybe you should make your bot follow one of th... [15:50:24] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2379915 (Antigng_) Most of my tasks don't generate such " unacceptable amount of traffic". [15:56:13] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2376097 (BBlack) >>! In T137707#2376258, @Antigng_ wrote: > My bot was using /w/index.php?action=raw to fetch the content of each page/redirect at zhwiki, then it... [16:10:42] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2379995 (Antigng_) Also, there doesn't exist a clear request rate limit for mediawiki api, as the rest api does. If you want to set one, you should document it. [16:11:12] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2379997 (jcrespo) For the API part, I would like to add that API infrastructure (application servers and databases) is specifically prepared to be separated from n... [16:22:54] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2380046 (Antigng_) I don't think api.php?action=query&prop=revisions&rvprop=content can be the same performant as index.php?action=raw, and the latter is the easie... [16:32:37] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2380159 (jcrespo) > I would appreciate it if there was a way to perform api.php?action=raw Please file a separate bug report for that. [17:09:51] DBA, Labs, Operations, Tool-Labs, Traffic: Antigng-bot improper non-api http requests - https://phabricator.wikimedia.org/T137707#2380275 (jcrespo) BTW, the API is definitely faster, one just need to use it efficiently: ``` $ time curl 'https://en.wikipedia.org/w/api.php?action=query&prop=r... [17:42:16] DBA, Labs, Labs-Infrastructure: Add shared edit rights between eranbot and copypatrol tools - https://phabricator.wikimedia.org/T137691#2380490 (kaldari) This is on the toolsdb. Probably doesn't need any work from you, @jcrespo. [21:41:49] DBA, Labs, Labs-Infrastructure: Add shared edit rights between eranbot and copypatrol tools - https://phabricator.wikimedia.org/T137691#2381148 (eranroz) This is the same DB as mentioned in T134392 (e.g s51306__copyright_p) and I agree to grant access to tool copypatrol (s53018) - including permissi...