The web in a box - a next generation web framework for the Perl programming language

IRC log for #mojo, 2017-11-11

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:11 jberger I have never been able to understand how to deal with purl's handling of I/my etc
00:11 jberger heh, check this out
00:11 jberger purl: sri's point
00:11 purl sri's point is that you can use a non-multipart request
00:11 jberger purl forget sri's point
00:11 purl jberger: I forgot sri's point
00:12 sri my point?
00:26 mohawk sri's point is that you should make websites with HTML, not JS
00:26 mohawk sri's point?
00:26 purl sri's point is that you should make websites with HTML, not JS
00:27 * mohawk cackles
00:28 batman jberger: guess I’m late to the party, but I don’t see any reason for why the chrome executable should be part of the url at all
00:29 jberger batman: you were the one that wanted the arguments in a url form :-P
00:29 jberger what good does that do if you can't also specify the executable?
00:32 batman Because the executable should just be found if host+port = “localhost”
00:32 sri batman: fix Mojo::Redis2!
00:33 jberger batman: and what do you do if it cannot or will not be found?
00:33 sri you used Mojo::Base::_monkey_patch
00:33 batman Environment variable, just like you have now
00:33 sri (╯°□°)╯︵ ┻━┻
00:33 Grinnz technically, some contributor used it in a pr and batman didnt fix it :P
01:03 disputin joined #mojo
01:21 aborazmeh joined #mojo
01:31 batman It’s very strange that I missed out on it. Commented on many other things...
01:34 batman at least i have proper internet now, so i can fix it!
01:34 Psiman \o/
01:36 Psiman As the contributor in question have been looking forward to getting this applied.. :)
01:59 dugword joined #mojo
02:04 disputin joined #mojo
02:06 batman Proper internet... yeah right 🙄
02:06 Grinnz mine just cut out too 🤕
02:07 batman At least I just managed to release mojo-mysql and mojo-redis2
02:07 * batman goes for a swim
02:07 Grinnz batman++
02:08 Psiman enjoy!
02:12 batman Thanks for all the help :)
02:12 kaare joined #mojo
02:13 Psiman No problem!
02:13 rickbol joined #mojo
03:04 Seth joined #mojo
03:18 noganex joined #mojo
05:04 dboehmer_ joined #mojo
05:29 dugword joined #mojo
06:59 Vandal joined #mojo
07:10 aborazmeh joined #mojo
08:35 dod joined #mojo
08:43 dod joined #mojo
10:36 geospeck joined #mojo
11:49 geospeck joined #mojo
11:50 geospeck joined #mojo
11:51 geospeck joined #mojo
12:57 sjn joined #mojo
13:09 PsimanX1 joined #mojo
13:14 syedi joined #mojo
13:16 syedi hi
13:16 purl hi, syedi.
14:21 gryphon joined #mojo
15:18 ChmEarl joined #mojo
15:37 geospeck joined #mojo
15:56 dod joined #mojo
16:22 geospeck joined #mojo
16:23 sri guess i'll bundle bootstrap, fontawesome, d3.js and epoch.js with minion for the admin ui
16:28 ghenry joined #mojo
16:31 maschine does it use websockets or https?
16:34 brunoramos joined #mojo
16:40 geospeck joined #mojo
16:40 geospeck joined #mojo
16:41 geospeck joined #mojo
16:42 geospeck_ joined #mojo
16:48 kaare joined #mojo
16:58 ghenry does mojo and emberjs work nicely together?
17:10 karjala_ joined #mojo
18:02 AndrewIsh joined #mojo
18:28 sri http, websockets wouldn't do much
18:29 sri i wonder if a short term history method would be useful in minion backends
18:30 sri extract(epoch from date_trunc('hour', created)), count(*) from minion_jobs group by 1 order by 1;
18:30 sri specifically a query like that
18:30 sri umm
18:30 sri select extract(epoch from date_trunc('hour', created)), count(*) from minion_jobs group by 1 order by 1;
18:31 sri to get the number of added jobs for every hour in the current queue content
18:32 sri which is 2 days by default
18:33 sri to get a history graph in the admin ui :)
18:34 sri basically a look back at the last two days with a graph showing how many jobs have been added every hour
18:35 dugword joined #mojo
18:54 geospeck joined #mojo
18:58 vicash joined #mojo
19:28 vicash hello. has anyone used Grafana with Mojolicious and/or Minion logs ? what kind of backend do you use for parsing the logs and loading into say InfluxDB ?  I would prefer to use a non-java solution or write my own in perl but if someone has used something in perl/python that works well out of the box with minimum configuration I would love to hear about it.
19:29 preaction i use ETL::Yertl to read the mysql database to produce Minion metrics using InfluxDB and Grafana
19:30 vicash preaction: I saw ETL::Yertl on metacpan as supporting InfluxDb but can I use Postgresql with it ?
19:30 preaction it uses DBI, so yes
19:30 vicash awesome
19:39 vicash preaction: as I understand the documentation for the yts tool, I can use ETL::Yertl::Adapter::influxdb as a module in my code to directly write to InfluxDB, avoiding the need to tail logs ?
19:41 preaction i wouldn't. i'm not guaranteeing those APIs yet. you probably want ygrok to parse the logs (why are you parsing logs?)
19:44 vicash well my minion jobs write some information on time elapsed and some other fields like number of results found in the logs.  i want to create Grafana dashboards around those fields to check what percentage of jobs took longer than usual and which jobs generated the most fields with results..
19:44 vicash all those things are written in the logs as of now and these logs are distributed across 32 minion workers. so i am trying to reduce my work and use Grafana and Influxdb to create some nice dashboards to have a managerial look at the data
19:45 preaction it would be better, i think, to write that data to the result of the Minion job, then it's right in the database. the start/finish time is already in the database
19:45 vicash i thought of that initially, but doesn't Minion remove the old finished jobs after some fixed time period ?
19:46 preaction but then you don't have to try parsing things, which is always fun
19:46 preaction it does, but that's why you grab it and move it into influxdb
19:46 vicash oh ok... that makes sense now
19:46 vicash cool
19:46 preaction and so you use ysql and yts to move the data from postgres to influx
19:47 vicash right.. all in a shell script !
19:47 preaction exactly!
19:47 vicash beautiful !
19:47 vicash thank you !
19:47 preaction but also, ygrok to parse the logs would also work. i just don't think it's good to do that if you don't have to (use it for apache / nginx logs where you can't get the data any other way)
20:32 sri next minion release will be 8.0
20:32 sri and i'll make a few breaking changes
20:32 sri backend->job_info($id) is going away
20:32 sri in favor of backend->list_jobs(0, 1, {ids => [$id]})
20:33 sri well technically also a ->[0]
20:33 sri same for worker_info and list_workers
20:34 sri has the advantage that it's a lot faster, less code, and allows new features in the admin ui :)
20:35 sri i want a minion job list view where i can filter by many different criteria
20:36 sri like a link to get to a list view of a jobs parents
20:36 sri /minion/jobs?id=1&id=2&id=3
20:37 sri perhaps i should just break ->list_jobs(*, *, {queue => 'foo'}) and friends too while i'm at it
20:37 sri so it can be {queues => ['foo', 'bar']}
20:38 sri {states => []}
21:21 sri guess the others are not as useful
21:37 disputin joined #mojo
22:22 karjala_ joined #mojo
22:25 karjala_ joined #mojo
22:33 dugword joined #mojo

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary