[12:28:15] phedenskog: FYI that's 3x more expensive than our current Kobiton budget (19k) [12:28:50] yeah :( [12:31:59] so who can host this at home? [12:32:59] I don't think it's impossible for us to get that budget approved [12:34:56] in my new house I'm sure I could find space for this, though... and the house is eligible for FTTH, might already have it [12:35:09] I'll put in the budget request and we'll see what happens [12:35:14] If we can have multiple Mac minis we can run desktop tests and skip some AWS instances. [12:39:18] please make a list of all the hardware we would need to buy, then [12:39:22] so we come up with our own figure [12:39:23] and I can have one Mac mini and one Android phone directly under my router :D Actually what we need is help from dpifke to get the limited wifi working so we can test and see what we get in an isolated environment. [12:39:32] yep [13:00:43] I did a quick calculation and the price for that setup if we buy it is $2000 (2 Mac minis, Five motos, 2 used iPhones). Then you need keyboard/mouse/monitor for initial setup and a usb hub. I'll create Google docs. [13:02:42] mac mini M1s cost less than $1000? [13:03:26] they do, wow. what a weird world we live in that it costs half of the top-of-the-line iphone [13:04:45] we'll probably want a raspberry pi or equivalent for the throttled wifi [13:05:07] temporary keyboard/mouse/monitor I already have if I set this up at my place [13:06:04] mac mini has 2 USB-A and 2 USB-C. if we have two of those, do we really need a USB hub? [13:06:27] they introduce issues, might as well avoid them if we can [13:25:17] No that could be enough. [13:25:32] Yes the new M1 is really cheap for what you get. [13:25:33] M1: MediaWiki Userpage - https://phabricator.wikimedia.org/M1 [15:19:24] Helpful bot is helpful [20:53:45] dpifke: so what's the price for a small office space where you live? [20:54:29] AaronSchulz: https://grafana.wikimedia.org/d/000000066/resourceloader?viewPanel=45&orgId=1&from=now-90d&to=now [20:54:38] I'm guessing that's buster-related :/ [21:08:22] indeed seems to match up both in terms of the timeline (biggest shift Feb 4-13), and in terms of latency (about +20ms) [21:08:24] links courtesy of legoktm and cdanis: https://w.wiki/34qu , https://phabricator.wikimedia.org/T273312#6788233 [22:08:40] phedenskog: Looks like $300/mo-ish, e.g. https://rapidcity.craigslist.org/off/d/rapid-city-beautiful-buell-building/7286398466.html [22:57:21] dpifke: I have a potential example of reqId differing between MW and plain PHP context [22:57:40] https://logstash.wikimedia.org/goto/0d4070b51524e9fa3663996e1d3a1052 [22:58:09] the same request has one reqId from MW perspective, but then in plain PHP context /etc/php-fatal-error it sees a different value [22:58:42] Hmm. [23:01:41] The one that MW sees is `X-bUnApAMNYAA0uKZekAAABH` [23:01:52] and later on plain PHP sees `79d7a047-c0ad-4355-9be5-4c6d375281e6` [23:02:30] https://github.com/wikimedia/puppet/blob/4f63bb2d46a46409bfa14295469bf85db725584e/modules/profile/files/mediawiki/php/php7-fatal-error.php#L149-L150 [23:02:56] $reqId = $_SERVER['HTTP_X_REQUEST_ID'] ?? $_SERVER['UNIQUE_ID'] ?? null; [23:03:01] vs [23:03:01] https://github.com/wikimedia/mediawiki/blob/7cc9b84af92f9cede9c4ce2522dcf238d6e76990/includes/WebRequest.php [23:03:04] $id = $_SERVER['HTTP_X_REQUEST_ID'] ?? $_SERVER['UNIQUE_ID'] ?? wfRandomString( 24 ); [23:03:15] Most peculiar