15:01:44 #startmeeting metrics team 15:01:45 Meeting started Fri Nov 29 15:01:44 2019 UTC. The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:45 Useful Commands: #action #agreed #help #info #idea #link #topic. 15:01:55 the riseup pads are down so maybe we continue in this nc pad 15:02:02 heh. 15:02:13 i sent you all a mail about the issues with the nc pad 15:02:16 it worked last time, what could go wrong. 15:02:17 but so far is working for us 15:02:20 :) 15:02:35 added a section to the meeting on interesting links to share :) 15:02:50 to the meeting pad* 15:03:29 anything to add to the topics? 15:03:37 not from me 15:04:02 okay, then let's start! 15:04:09 collector disk space, and unreliable physical host (irl) 15:04:25 we are are 80% disk space and generating a lot of alerts 15:04:30 we have the option to double the disk space 15:04:43 let's do it! 15:04:44 i've noticed though the host has gone down a couple of times 15:04:49 and maybe it is a good idea to move it 15:05:00 do we have anything that tells us if it's missing data when it's gone down? 15:05:02 do you know why it went down? 15:05:07 i think it is the vm host being rebooted 15:05:33 we have the reference checker that tells us if things go really bad. 15:05:55 but I think in most cases it would just sync over things from the other instance. 15:06:00 it might be that the reboots happen quickly enough that everything is ok 15:06:33 I can't say for sure that the syncing works, but I didn't see any breakage. 15:06:47 ok then, i will just ask for the disk space to be doubled 15:06:54 that sounds great! 15:06:57 that's all on this topic 15:07:09 in the future I'd want collector modules to clean up better after themselves. 15:07:28 just like they're cleaning up in the recent/ directories. they should do the same in out/ directories. 15:07:47 but you know, they still need to grow older until they understand the benefits of cleaning up.... 15:07:56 it's on the list. 15:08:05 heh 15:08:24 #32644 15:08:24 okay, cool! 15:08:31 nice, that was quick. 15:08:42 "metrics-team" keyword on IRC (karsten) 15:08:55 this is based on a suggestion by the network team. 15:09:03 they have a similar keyword: "network-team". 15:09:15 and one for the anti cen sor ship team 15:09:16 team members have a highlight for that. 15:09:29 ah, cool. 15:09:37 +1 for the metrics-team keyword 15:09:44 it sounds like a good idea to me. I already have a highlight for "metrics" from years ago. 15:10:03 i used to have highlights for metrics and atlas but i think i don't right now 15:10:21 would you want to create one for metrics-team? 15:10:28 i could do, yes 15:10:43 cool! I'll do the same. 15:11:00 do we have an index of these team highlights? 15:11:01 gaba, are you collecting team highlights somewhere? 15:11:10 and when and when not to use them 15:11:13 i'm not doing it right now 15:11:22 we could add this info to each team wiki page 15:11:25 i imagine network team has a lot of people on it, and they all get distracted 15:11:34 something about amplification attacks 15:11:45 that is a good idea 15:11:51 we can add some information on when to use it 15:12:04 in general is for when we need to reach anybody in the team 15:12:06 I can update the metrics team page. 15:12:13 ty 15:12:48 should we do more? 15:13:09 like tell the network team who initiated this idea? 15:13:28 it doesn't reach anybody 15:13:32 it reaches everybody 15:13:57 that's true. 15:14:44 but that works with mailing lists, too. 15:15:14 i treat an irc highlight as a lot more urgent than an email 15:15:25 yes 15:15:30 and less urgent than a signal msg 15:15:36 nope, the same 15:15:43 i get irc highlights on my phone (when it's working) 15:15:49 oh, i see 15:15:50 oh, I don't. 15:15:59 i do not have irc in my phone 15:16:35 i could see if i can come up with a weechat script that enables/disables the highlight depending on when i'm paying attention to irc 15:16:55 so i won't get highlights when i'm not working 15:17:08 but then i wouldn't see them in backlog (which i guess is not the aim of the highlight) 15:17:31 we can think about this more and make a decision next week. 15:17:44 ok 15:18:06 okay, moving on: 15:18:07 Pending reviews (karsten) 15:18:20 just wanted to ask whether these are on your radar. 15:18:39 these = #25924 and #32605 15:18:56 yes they are, but it's been a 3 day week this week and i wanted to get the exit scanner running over the weekend to make sure it's stable 15:19:05 sounds good! 15:19:29 they will probably be monday 15:19:43 cool! 15:19:46 roadmap? 15:19:55 yep 15:20:08 ok 15:20:50 how is the exit scanner going? 15:20:52 I assume I should leave #32473 where it is? 15:21:54 yes, we will want to check published times are matching up and also detail the exact differences in documentation 15:22:14 differences regarding ports? 15:22:17 yes, we might want to give warning to exit operators about the change 15:22:56 so, did you deploy it to run over the weekend? 15:23:22 i did, on aws 15:23:39 great! happy to look at the data again after the weekend. 15:23:44 \o/ 15:24:14 we can leave it running until tuesday, if i do reviews on monday 15:24:21 and then i'll give you the data tuesday morning first thing 15:24:21 sure. 15:25:26 back to the roadmap. 15:25:41 I didn't get to #32126 yet. will keep it. 15:26:04 #25924 is under review (on monday). 15:26:41 ah, we need to deploy a new onionoo-backend-01 next week. 15:27:29 speaking of, could we just deploy -01 and -02 and disable omeiense and oo-hetzner-03? 15:27:42 i think my roadmap items look the same next week, but will hopefully be all for review and/or done by the end of the week, i will need to make new tickets for the check integration and dns server 15:27:49 having three different setups doesn't make it easier to update things. 15:27:54 yes, we could request a new backend be provisioned 15:28:06 the spec of the machine worked well 15:28:09 can you do that, with the same spec? 15:28:17 yes 15:28:26 thanks! 15:28:39 and shall we meet on tuesday to do the setup? 15:29:49 and did you ask for onionoo-backend-01 to be decomissioned and recreated? 15:29:56 or do we not have to do that? 15:30:03 turns out we can do it all ourselves 15:30:10 to remove something from varnish, we just turn off the webserver 15:30:11 ah, okay. 15:30:12 it's automatic 15:30:19 yes, I saw that. pretty cool! 15:30:22 to put it back, turn it back on (: 15:30:38 and we just purge the relevant dirs on -01 ourselves? 15:30:49 and recreate them? 15:31:08 yeah 15:31:15 cool. 15:32:00 so, tuesday? 15:32:15 yes, sounds good 15:32:19 1500? 15:32:23 ok 15:32:54 great! 15:33:25 anything else on the roadmap topic? 15:33:35 not from me 15:34:13 okay! final topic: 15:34:14 [announcement] we are testing phase of migration to gitlab - it would be good to get people to check for possible problems https://pad.riseup.net/p/gitlab-migration-problems 15:34:31 I already looked and gave ahf some feedback via email, because the pad didn't work. 15:35:09 happy to see this migration making progress! 15:35:21 thanks 15:35:54 something else to add to today is a reminder that we may do some work for the new data portal that simply secure is going to work on (thanks antonela for the reminder). I will do a follow up with them in january 15:36:05 ok cool 15:36:07 * antonela o/ 15:36:22 great! 15:37:08 alright, any more topics for today? 15:37:16 not from me 15:37:24 nop 15:37:39 yay. have a great weekend, everyone! 15:37:50 #endmeeting