--- Log opened vie ago 05 00:00:12 2022 --- Log closed vie ago 05 02:25:21 2022 --- Log opened vie ago 05 02:25:31 2022 02:25 -!- Irssi: #friendica: Total of 9 nicks [0 ops, 0 halfops, 0 voices, 9 normal] 02:25 -!- Irssi: Join to #friendica was synced in 18 secs 07:42 < fikabot_> ๐Ÿ’ฌ should this be in the official admin documentation? probably should wait for https://github.com/friendica/friendica/pull/11702 to be finished though before adding it. 08:23 < fikabot_> ๐Ÿ’ฌ s_t_e_v_e: what is "this" in that sentence? 08:24 < fikabot_> ๐Ÿ’ฌ your writeup 08:24 < fikabot_> ๐Ÿ’ฌ oh 08:24 < fikabot_> ๐Ÿ’ฌ Yeah that and a cut down version of my hand holding install guide I think should be in the main documentation 08:24 < fikabot_> ๐Ÿ’ฌ https://nequalsonelifestyle.com/2022/07/30/creating-friendica-server-ubuntu/ 09:25 < fikabot_> ๐Ÿ’ฌ btw the weird thing is that after I've switched to daemon from cron Message queues were dropping down by the thousands daily. i reached some 58k from 150k. and then I did your latest tutorial with daemon and systemd and it wont go bellow 58k haha. all works fine system wide but no clue what the hell ๐Ÿ˜„ 09:46 < fikabot_> ๐Ÿ’ฌ That sounds like correlation is not causation principle :) 09:47 < fikabot_> ๐Ÿ’ฌ Since the system is essentially the same it's just that systemd is making sure if the process dies it is revived 09:47 < fikabot_> ๐Ÿ’ฌ but double check that it is still running etc. 09:48 < fikabot_> ๐Ÿ’ฌ yeah is running i checked...and there are new Message queues coming and new are get rid off....so it burns down hundreds a day probably 09:48 < fikabot_> ๐Ÿ’ฌ but weird that got stuck at 58k and before that was burning them like nuts ๐Ÿ˜„ 09:52 < fikabot_> ๐Ÿ’ฌ Part of what I'd like to do is keep statistics about the worker history 09:52 < fikabot_> ๐Ÿ’ฌ by type and time 09:53 < fikabot_> ๐Ÿ’ฌ for that dashboard 09:53 < fikabot_> ๐Ÿ’ฌ how many of each type per hour, min, max, median, and mean time for completion per hour etc 09:53 < fikabot_> ๐Ÿ’ฌ I'm an engineer I love my data 09:54 < fikabot_> ๐Ÿ’ฌ but first simple stuff like current task queue statistics so can see how many of each work type there is 09:57 < fikabot_> ๐Ÿ’ฌ Could it happen that certain jobs got stuck in the queue that ran into an unexpected error? 10:01 < fikabot_> ๐Ÿ’ฌ or those really big queries that got stuck at the database level have reared their head again 10:04 < fikabot_> ๐Ÿ’ฌ That would also sound plausible. It would be nice to keep account of how many times a given job was retried, how much accumulated CPU/DB time it has used up to now, etc. 10:04 < fikabot_> ๐Ÿ’ฌ yes that too to add to a stats table 10:06 < fikabot_> ๐Ÿ’ฌ In general, this would sound like something counterproductive by loading the already loaded database even more, however if you consider how many jobs you need to process in a day, it's not that grave and it might even be controlled by a boolean setting. 10:06 < fikabot_> ๐Ÿ’ฌ what would? 10:07 < fikabot_> ๐Ÿ’ฌ It should be pretty low cost to update stats variables on a once a minute basis or something and only at 1 hour granularity 10:07 < fikabot_> ๐Ÿ’ฌ before that I'd just make a more informative dashboard with present data 10:15 < fikabot_> ๐Ÿ’ฌ Oh, you would store some kind of `top` on the mariadb processes? 10:36 < fikabot_> ๐Ÿ’ฌ I was thinking each worker would write to raw stats table and then periodically a stat summary process would come through and vacuum up anything more than an hour old, calculate the stats, and then write that to a stats summary table that could be used on a dashboard 10:36 < fikabot_> ๐Ÿ’ฌ I'm literally just coming up with this off the top of my head though --- Log closed vie ago 05 16:43:32 2022 --- Log opened vie ago 05 16:43:41 2022 16:43 -!- Irssi: #friendica: Total of 9 nicks [0 ops, 0 halfops, 0 voices, 9 normal] 16:43 -!- Irssi: Join to #friendica was synced in 19 secs 18:19 -!- kernelpanic is now known as fatalerrors --- Log closed sรกb ago 06 00:00:13 2022