[01:10] Hmm. [01:10] Only 583 timeouts. [01:10] Perhaps the lack of index bloat is useful. === _mup__ is now known as _mup_ [09:31] Yippie, build fixed! [09:31] Project devel build #894: FIXED in 5 hr 41 min: https://lpci.wedontsleep.org/job/devel/894/ === _mup__ is now known as _mup_ [21:43] mtaylor: so [21:43] mtaylor: when are we getting those blueprints patches ? :) [21:44] heh. hehehehe [21:44] lifeless: which ones was I going to do again? [21:44] I dunno [21:44] and wait - aren't you meging them with bugs? [21:44] you had some openstack wishlist thingies [21:44] yeah, and ? [21:45] well, I finally have another guy working with me [21:45] oh oh oh [21:45] who is doing some hacking - although he's heads down in jenkins and gerrit right now [21:45] did you see we got the bugsummary fact/dimension setup going ? [21:45] BUT - we should be able to get to some of these sorts of things eventually now (bandwidth is good) [21:45] I didn't - what do those words mean? [21:45] data warehousing approach [21:46] a fact table - a table with (0,1) rows per unique combination of dimensions (in this case bugtasktarget, status, importance, has-patch, fixed-upstream, viewable-by) [21:46] lovely! [21:47] and then you slice n dice the table to pull out stats [21:47] it has a column containing the counter for how many tasks matched that point in the table space [21:48] e.g. on https://bugs.launchpad.net/ubuntu/ [21:48] the two portlets - tags and the counts [21:48] used to take 10+ seconds each to render [21:48] now the tags one is subsecond and the count one is 1.3 seconds 99% of the time, and not over 2s ever in the stats I've looked at [21:49] oh, tag is another dimension [21:50] excellent! [21:50] nicely done [21:50] in the last 6 months we've dropped out 99% render time site wide [21:50] ... by 50% [22:20] lifeless: you know what - puppet is great except when it's not [22:26] orly? software not being great? [22:31] I'm shocked