*** markvoelker has joined #openstack-tc | 00:23 | |
*** markvoelker has quit IRC | 00:28 | |
*** jamesmcarthur has joined #openstack-tc | 00:29 | |
*** nicolasbock has quit IRC | 00:29 | |
*** zhurong has quit IRC | 00:40 | |
*** tetsuro has quit IRC | 01:05 | |
*** jamesmcarthur has quit IRC | 01:07 | |
*** wxy-xiyuan_ has joined #openstack-tc | 01:07 | |
*** jamesmcarthur has joined #openstack-tc | 01:07 | |
*** zhurong has joined #openstack-tc | 01:08 | |
*** jamesmcarthur has quit IRC | 01:12 | |
*** tetsuro has joined #openstack-tc | 01:16 | |
*** markvoelker has joined #openstack-tc | 01:26 | |
*** markvoelker has quit IRC | 01:37 | |
*** markvoelker has joined #openstack-tc | 01:37 | |
*** jamesmcarthur has joined #openstack-tc | 01:38 | |
*** markvoelker has quit IRC | 01:42 | |
*** markvoelker has joined #openstack-tc | 02:02 | |
*** tetsuro has quit IRC | 02:14 | |
*** jamesmcarthur has quit IRC | 02:16 | |
*** tetsuro has joined #openstack-tc | 02:17 | |
*** wxy-xiyuan_ is now known as wxy-xiyuan | 02:33 | |
*** jamesmcarthur has joined #openstack-tc | 02:36 | |
*** jamesmcarthur has quit IRC | 02:41 | |
*** ricolin has joined #openstack-tc | 02:47 | |
*** tetsuro has quit IRC | 03:37 | |
*** tetsuro has joined #openstack-tc | 03:38 | |
*** markvoelker has quit IRC | 03:49 | |
*** fungi has quit IRC | 04:21 | |
*** fungi has joined #openstack-tc | 04:26 | |
*** bnemec has quit IRC | 05:19 | |
*** markvoelker has joined #openstack-tc | 05:50 | |
*** markvoelker has quit IRC | 05:55 | |
*** jaosorior has quit IRC | 06:41 | |
*** tetsuro has quit IRC | 07:16 | |
*** tetsuro has joined #openstack-tc | 07:16 | |
*** iurygregory has joined #openstack-tc | 07:23 | |
*** tosky has joined #openstack-tc | 07:24 | |
*** jaosorior has joined #openstack-tc | 07:24 | |
evrardjp | o/ | 07:37 |
---|---|---|
openstackgerrit | Merged openstack/governance master: check-review-status: update age calc https://review.opendev.org/684784 | 07:40 |
*** lpetrut has joined #openstack-tc | 07:45 | |
ricolin | o/ | 07:47 |
*** e0ne has joined #openstack-tc | 08:31 | |
*** jaosorior has quit IRC | 08:32 | |
*** e0ne has quit IRC | 08:52 | |
evrardjp | I will miss part of the office hours, sorry for that | 08:52 |
evrardjp | o/ | 08:59 |
ttx | o/ | 09:02 |
ttx | office hours! woohoo! | 09:02 |
evrardjp | wooot | 09:02 |
evrardjp | how was your nordic trip ? | 09:03 |
ttx | It was good! Met with some interesting people trying to build more sustainable infrastructure | 09:03 |
ttx | I have one topic to discuss... telemetry! | 09:03 |
ttx | There are growing concerns with the state of that stack, especially its dependency on Gnocchi, which appears to be unmaintained | 09:04 |
ttx | as hinted at the bottom of https://github.com/gnocchixyz/gnocchi/pull/1047 | 09:05 |
ttx | We had some discussion of converging Monasca and Telemetry back in April, but I don't think that went anywhere | 09:06 |
ttx | Thoughts? | 09:06 |
ricolin | o/ | 09:09 |
ricolin | ttx another issue with Telemetry is Aodh didn't sync up with Gnocchi that makes cpu_util broken since Stein | 09:13 |
ricolin | see http://lists.openstack.org/pipermail/openstack-discuss/2019-August/008541.html | 09:13 |
*** e0ne has joined #openstack-tc | 09:13 | |
ricolin | witek is working on improve Monasca to cover most use cases we been using Telemetry like self-healing with Monasca or installs and enables per default libvirt monitoring in Monasca DevStack plugin | 09:16 |
ttx | ok | 09:17 |
ricolin | Monasca can already cover Autoscaling use cases | 09:19 |
lxkong | ricolin: just to clarify, problem is not 'Aodh didn't sync up with Gnocchi', truth is, people just find it very difficult to implement the same function as previous ceilometer 'cpu_util' in Gnocchi. | 09:19 |
ricolin | lxkong, sorry for the poor choose of words. Thanks for help that out. We're on the same page:) | 09:21 |
ricolin | lxkong, the new Gnocchi cpu metering is also not implemented in Aodh too | 09:22 |
evrardjp | ttx: would focusing on prometheus exporters for some software work for people? | 09:23 |
lxkong | Gnocchi is hard | 09:23 |
evrardjp | because it decouples the problem differently | 09:23 |
ricolin | lxkong, is your team test on any other monitoring tools? | 09:23 |
ricolin | like Monsca or prometheus? | 09:24 |
lxkong | nope, we are using Icinga for internal system monitoring. We are using Ceilometer for metering, Aodh for notification | 09:24 |
lxkong | we don't use Gnocchi | 09:25 |
lxkong | our concerns for Monasca is it's using Kafka and Time Series Databases(accoding to the doc) which are not deployed in our stack | 09:25 |
evrardjp | lxkong: what's in your stack? | 09:26 |
ttx | evrardjp: I feel like weaknesses in the stack are worth discussing at the TC session at the PTG | 09:27 |
evrardjp | ttx: agreed | 09:27 |
ricolin | ttx definitely | 09:27 |
lxkong | evrardjp: all our services are using rabbitmq and mysql, deploying kafka and Time Series Database dedicated for Monasca sounds a bit scary :-) | 09:27 |
ttx | we have a number of tools at our disposal to push in a given direction, but there is no clear direction on some of those | 09:27 |
evrardjp | lxkong: I get that :) | 09:28 |
ricolin | lxkong, for performance concerns or stability? | 09:28 |
ricolin | or upgrade? | 09:28 |
ttx | ricolin: complexity I'd say | 09:28 |
lxkong | expertise | 09:28 |
ricolin | lxkong, got it | 09:29 |
lxkong | well, sorry for chiming in | 09:30 |
* lxkong goes to bed | 09:31 | |
*** iurygregory has quit IRC | 09:31 | |
ricolin | ttx I suggest the PTG session is not only for maintain Telemetry but also discuss about overall Metering (and it's long term maintain) across OpenStack | 09:31 |
ricolin | lxkong, NP | 09:31 |
ttx | sure. taking a step back | 09:31 |
ricolin | lxkong, will ask witek to bothering you while you asleep! | 09:31 |
*** jaosorior has joined #openstack-tc | 09:44 | |
*** iurygregory has joined #openstack-tc | 09:52 | |
*** markvoelker has joined #openstack-tc | 09:53 | |
*** markvoelker has quit IRC | 09:58 | |
*** tetsuro has quit IRC | 10:06 | |
*** tetsuro has joined #openstack-tc | 10:06 | |
*** tetsuro has quit IRC | 10:11 | |
*** tetsuro has joined #openstack-tc | 10:13 | |
*** jaosorior has quit IRC | 10:18 | |
*** tetsuro_ has joined #openstack-tc | 10:20 | |
*** tetsuro has quit IRC | 10:21 | |
asettle | o/ | 10:34 |
*** dklyle has quit IRC | 10:49 | |
*** dklyle has joined #openstack-tc | 10:49 | |
jroll | lxkong: don't be sorry, we appreciate you chiming in :) | 11:01 |
jroll | ++ for discussing metering, we've put off upgrading the telemetry stack due to these unknowns | 11:02 |
*** nicolasbock has joined #openstack-tc | 11:07 | |
*** jamesmcarthur has joined #openstack-tc | 11:15 | |
*** jamesmcarthur has quit IRC | 11:19 | |
ttx | you're not the only ones | 11:43 |
*** tetsuro_ has quit IRC | 11:52 | |
*** tetsuro has joined #openstack-tc | 11:52 | |
*** markvoelker has joined #openstack-tc | 11:54 | |
*** markvoelker has quit IRC | 11:59 | |
evrardjp | I have the impression we are attending the birth of telemetry 2.0 | 12:02 |
*** markvoelker has joined #openstack-tc | 12:03 | |
*** jaosorior has joined #openstack-tc | 12:08 | |
*** jaosorior has quit IRC | 12:08 | |
*** jaosorior has joined #openstack-tc | 12:08 | |
*** tetsuro has quit IRC | 12:20 | |
gmann | o/ | 12:24 |
*** jamesmcarthur has joined #openstack-tc | 12:26 | |
*** tetsuro has joined #openstack-tc | 12:26 | |
*** jamesmcarthur has quit IRC | 12:31 | |
*** ijolliffe has joined #openstack-tc | 12:45 | |
* mnaser is in the school of thought of "lets just let ceilometer move on and use the newer tool that have emerged" | 12:50 | |
mnaser | i.e. prom | 12:50 |
*** zhipeng has joined #openstack-tc | 12:50 | |
*** jamesmcarthur has joined #openstack-tc | 12:51 | |
tosky | prom? | 12:57 |
jroll | prometheus | 12:58 |
*** mriedem has joined #openstack-tc | 12:59 | |
*** lpetrut has quit IRC | 13:08 | |
dhellmann | yeah, this doesn't seem like an area where we want to reinvent just because we could | 13:13 |
dhellmann | it turns out that metrics a hard problem to solve at any real scale | 13:13 |
mugsie | yeah - metrics + tracing I am of the opinion we should conform to the emerging standards (prometheous + jeager / opentracing) | 13:28 |
dhellmann | ++ | 13:31 |
evrardjp | agreed | 13:31 |
cloudnull | o/ | 13:41 |
*** bnemec has joined #openstack-tc | 13:43 | |
ttx | I tend to agree but I'm not seeing any significant progress in that direction either | 13:45 |
mugsie | I think it is something we would have to drive | 13:45 |
smcginnis | Cycle goal? | 13:47 |
mugsie | I think multi cycle goal | 13:48 |
mugsie | maybe a sig / pop up team | 13:48 |
mugsie | they would need to produce the colateral for usage by teams, and then educate the teams why | 13:49 |
mugsie | some tried in nova this cycle (or maybe last) and I don't think it went well | 13:49 |
zaneb | I don't disagree, but the job of Telemetry in OpenStack is to produce *userspace* metrics/alarms in a multi-tenant way | 13:51 |
zaneb | I would humbly suggest that the word "just" doesn't have a place in this conversation ;) | 13:52 |
mugsie | zaneb: sure - the multi tenant thing is more interesting - but I would suggest building on top of the current solutions, not inventing our own. afaik most of the tools even support gnochi as a data store now | 13:57 |
mugsie | and no, "just" is understating it a bit | 13:57 |
zaneb | +1 for building on top of prometheus | 13:58 |
mugsie | I can even think of a few ways we could do multi tenant, but would need to test and prove it out first | 13:59 |
jroll | something like "magnum for prometheus"? | 14:00 |
mugsie | I was thinking of a centralised one | 14:01 |
*** diablo_rojo has joined #openstack-tc | 14:01 | |
mriedem | mugsie: nova re https://review.opendev.org/#/c/640688/ | 14:02 |
mugsie | mriedem: thanks, I was lookig for that | 14:03 |
mugsie | honestely, dansmith 's last comment sums up my views | 14:04 |
mugsie | (I know the -2 was procedural, not a block) | 14:04 |
*** jamesmcarthur has quit IRC | 14:06 | |
*** jamesmcarthur has joined #openstack-tc | 14:37 | |
coreycb | hi all, please could we get a wider vote on https://review.opendev.org/#/c/686253/? it's needed for the train cycle. | 14:48 |
coreycb | it's a rollcall-vote | 14:48 |
asettle | tc-members ^ | 14:49 |
jungleboyj | Thanks. Will take a look. | 14:49 |
jroll | easy enough :) | 14:52 |
evrardjp | probably need more eyes on https://review.opendev.org/#/c/686251/ too | 14:53 |
mugsie | aye - that needs to merge first | 14:54 |
njohnston | I guess the Depends-On isn't working since the linter job in 686253 is breaking because it doesn't see the change from 686251 | 14:55 |
mugsie | it won't for the project config repo afaik - due to the special status of the repo | 14:55 |
jroll | correct | 14:55 |
njohnston | Ah! Good to know. | 14:55 |
fungi | yes, openstack/project-config is marked as being safe to share secrets from, so zuul won't allow speculative execution of proposed changes for it | 14:56 |
fungi | at least not in pre-review pipelines like check | 14:57 |
coreycb | you all are fast, thanks for the quick attention | 15:13 |
*** jamesmcarthur_ has joined #openstack-tc | 15:23 | |
*** markvoelker has quit IRC | 15:24 | |
*** jamesmcarthur has quit IRC | 15:26 | |
*** iurygregory has quit IRC | 15:31 | |
*** e0ne has quit IRC | 15:55 | |
*** kmalloc has left #openstack-tc | 16:43 | |
*** ricolin has quit IRC | 17:03 | |
*** markvoelker has joined #openstack-tc | 17:25 | |
*** ijolliffe has quit IRC | 17:26 | |
*** markvoelker has quit IRC | 17:29 | |
*** ijolliffe has joined #openstack-tc | 17:33 | |
*** ijolliffe has quit IRC | 17:56 | |
*** jamesmcarthur_ has quit IRC | 18:02 | |
*** jamesmcarthur has joined #openstack-tc | 18:02 | |
*** markvoelker has joined #openstack-tc | 18:06 | |
*** ijolliffe has joined #openstack-tc | 18:06 | |
*** jamesmcarthur has quit IRC | 18:14 | |
*** mriedem has quit IRC | 18:31 | |
*** mriedem has joined #openstack-tc | 18:32 | |
*** ijolliffe has quit IRC | 18:46 | |
*** markvoelker has quit IRC | 18:50 | |
*** dhellmann has quit IRC | 18:51 | |
*** dhellmann has joined #openstack-tc | 18:51 | |
*** markvoelker has joined #openstack-tc | 18:54 | |
*** ianw_pto is now known as ianw | 18:59 | |
openstackgerrit | Nate Johnston proposed openstack/governance master: Post TC-election info-sheet https://review.opendev.org/683432 | 19:05 |
*** ijolliffe has joined #openstack-tc | 19:08 | |
*** jamesmcarthur has joined #openstack-tc | 19:17 | |
*** ijolliffe has quit IRC | 19:18 | |
openstackgerrit | Amy Marrich (spotz) proposed openstack/governance-sigs master: New UC Rep https://review.opendev.org/687392 | 19:18 |
*** ijolliffe has joined #openstack-tc | 19:26 | |
*** ijolliffe has quit IRC | 19:34 | |
*** jamesmcarthur has quit IRC | 19:52 | |
*** jamesmcarthur has joined #openstack-tc | 19:53 | |
*** jamesmcarthur has quit IRC | 19:58 | |
*** jamesmcarthur has joined #openstack-tc | 20:15 | |
*** mtreinish has quit IRC | 20:21 | |
*** e0ne has joined #openstack-tc | 20:23 | |
*** e0ne has quit IRC | 20:28 | |
*** mtreinish has joined #openstack-tc | 21:00 | |
dhellmann | mugsie: regarding centralized metrics, https://github.com/openstack/ironic-prometheus-exporter has a somewhat interesting approach but I think we could do better by actually instrumenting services | 21:14 |
*** nicolasbock has quit IRC | 21:14 | |
*** markvoelker has quit IRC | 21:43 | |
*** jamesmcarthur has quit IRC | 22:04 | |
*** ijolliffe has joined #openstack-tc | 22:06 | |
*** jamesmcarthur has joined #openstack-tc | 22:35 | |
*** jamesmcarthur has quit IRC | 22:36 | |
*** jamesmcarthur has joined #openstack-tc | 22:36 | |
*** jamesmcarthur has quit IRC | 22:39 | |
*** jamesmcarthur has joined #openstack-tc | 22:45 | |
*** jamesmcarthur has quit IRC | 22:50 | |
*** tosky has quit IRC | 23:10 | |
*** diablo_rojo has quit IRC | 23:17 | |
*** markvoelker has joined #openstack-tc | 23:32 | |
*** mriedem has quit IRC | 23:51 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!