| *** flwang has quit IRC | 00:27 | |
| openstackgerrit | Samuel Merritt proposed a change to openstack/ceilometer: Fix handling of bad paths in Swift middleware. https://review.openstack.org/45983 | 00:33 |
|---|---|---|
| *** nosnos has joined #openstack-metering | 00:36 | |
| *** flwang has joined #openstack-metering | 01:18 | |
| openstackgerrit | Samuel Merritt proposed a change to openstack/ceilometer: Fix handling of bad paths in Swift middleware https://review.openstack.org/45983 | 01:26 |
| *** terriyu has quit IRC | 01:31 | |
| *** haomeng has quit IRC | 01:45 | |
| *** d34dh0r53 has joined #openstack-metering | 01:59 | |
| *** d34dh0r53 has quit IRC | 02:01 | |
| *** yanghy has joined #openstack-metering | 02:13 | |
| *** SergeyLukjanov has joined #openstack-metering | 02:13 | |
| *** yanghy has left #openstack-metering | 02:22 | |
| *** Ruetobas has joined #openstack-metering | 02:29 | |
| *** yanghy has joined #openstack-metering | 02:30 | |
| *** anteaya_ has joined #openstack-metering | 02:31 | |
| *** anteaya has quit IRC | 02:32 | |
| *** lnxnut has joined #openstack-metering | 02:33 | |
| *** Ruetobas has quit IRC | 02:35 | |
| *** Ruetobas has joined #openstack-metering | 02:35 | |
| *** anteaya_ has quit IRC | 02:37 | |
| *** Ruetobas has quit IRC | 02:43 | |
| openstackgerrit | Yuuichi Fujioka proposed a change to openstack/python-ceilometerclient: Add check necessary parameter https://review.openstack.org/45989 | 02:58 |
| *** SergeyLukjanov has quit IRC | 03:10 | |
| *** lnxnut has quit IRC | 03:17 | |
| *** Guest11685 has quit IRC | 03:18 | |
| *** lillie has joined #openstack-metering | 03:19 | |
| *** lillie is now known as Guest31783 | 03:19 | |
| *** herndon_ has joined #openstack-metering | 03:40 | |
| *** shang has joined #openstack-metering | 03:43 | |
| *** herndon_ has quit IRC | 04:02 | |
| *** evanjfraser has quit IRC | 04:06 | |
| *** gordc has joined #openstack-metering | 04:19 | |
| *** shang has quit IRC | 04:29 | |
| *** shang has joined #openstack-metering | 04:31 | |
| *** gordc has quit IRC | 04:44 | |
| *** boris-42 has joined #openstack-metering | 04:51 | |
| openstackgerrit | Jenkins proposed a change to openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/45175 | 06:02 |
| *** fnaval_ has quit IRC | 07:08 | |
| *** boris-42 has quit IRC | 07:35 | |
| *** eglynn has joined #openstack-metering | 07:57 | |
| *** flwang1 has joined #openstack-metering | 08:03 | |
| *** flwang has quit IRC | 08:06 | |
| *** fc__ has quit IRC | 08:06 | |
| *** flwang has joined #openstack-metering | 08:14 | |
| *** flwang1 has quit IRC | 08:16 | |
| *** lexx has joined #openstack-metering | 08:19 | |
| *** lexx has quit IRC | 08:27 | |
| *** lsmola has quit IRC | 08:35 | |
| *** fc__ has joined #openstack-metering | 08:43 | |
| openstackgerrit | Julien Danjou proposed a change to openstack/ceilometer: run-tests: fix MongoDB start wait https://review.openstack.org/45642 | 08:43 |
| openstackgerrit | Julien Danjou proposed a change to openstack/ceilometer: Include auth_token middleware in sample config https://review.openstack.org/45886 | 08:46 |
| openstackgerrit | Julien Danjou proposed a change to openstack/ceilometer: Update config generator https://review.openstack.org/45885 | 08:46 |
| *** boris-42 has joined #openstack-metering | 08:49 | |
| *** lsmola has joined #openstack-metering | 08:50 | |
| *** mmcardle has joined #openstack-metering | 08:52 | |
| *** mmcardle has quit IRC | 08:54 | |
| *** mmcardle has joined #openstack-metering | 08:54 | |
| *** mmcardle has quit IRC | 08:58 | |
| *** lsmola has quit IRC | 08:59 | |
| *** mmcardle has joined #openstack-metering | 09:07 | |
| *** Alexei_987 has joined #openstack-metering | 09:08 | |
| Alexei_987 | jd__: Hello | 09:08 |
| jd__ | Alexei_987: hi | 09:09 |
| Alexei_987 | jd__: wanted to ask do we still have issues with Mongo connection pool? | 09:09 |
| Alexei_987 | jd__: I'm trying to review storage related issues that we have | 09:10 |
| jd__ | Alexei_987: not as far as I know | 09:11 |
| jd__ | Alexei_987: what do your errors look like? | 09:11 |
| Alexei_987 | jd__: on master branch it's ok for me. I have failures on my branch and I guess it's caused by my changes | 09:12 |
| Alexei_987 | jd__: just want to be sure that it's ok for everyone | 09:12 |
| jd__ | ack | 09:13 |
| *** lsmola has joined #openstack-metering | 09:14 | |
| *** flwang has quit IRC | 09:23 | |
| *** flwang has joined #openstack-metering | 09:24 | |
| *** lexx has joined #openstack-metering | 09:34 | |
| *** Alexei_987 has quit IRC | 09:39 | |
| openstackgerrit | A change was merged to openstack/python-ceilometerclient: Fix a typo in "sample-create" help message https://review.openstack.org/45702 | 09:40 |
| *** ruhe has joined #openstack-metering | 09:46 | |
| *** mmcardle has quit IRC | 09:59 | |
| *** flwang has quit IRC | 10:03 | |
| *** ruhe has quit IRC | 10:03 | |
| *** mmcardle has joined #openstack-metering | 10:06 | |
| *** ruhe has joined #openstack-metering | 10:17 | |
| openstackgerrit | Cyril Roelandt proposed a change to openstack/python-ceilometerclient: Help messages: specify which options are required https://review.openstack.org/45870 | 10:25 |
| *** mmcardle has quit IRC | 10:40 | |
| *** mmcardle has joined #openstack-metering | 10:42 | |
| *** blinky_ghost has joined #openstack-metering | 10:48 | |
| openstackgerrit | Mehdi Abaakouk proposed a change to openstack/ceilometer: Don't publish samples if resource_id in missing https://review.openstack.org/46048 | 10:49 |
| openstackgerrit | Mehdi Abaakouk proposed a change to openstack/ceilometer: Print the exception of samples publishing https://review.openstack.org/46049 | 10:49 |
| blinky_ghost | Hi guys, one quick question that I couldn't find on documentation. Does ceilometer support billing and Is there some guide to implement it? thanks | 10:51 |
| EmilienM | blinky_ghost: ceilometer project aims to deliver a unique point of contact for billing systems to acquire all of the measurements but is not a billing system. | 11:06 |
| blinky_ghost | EmilienM: so it only works like an API right? | 11:08 |
| *** flwang has joined #openstack-metering | 11:16 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Add testing of migrations from oslo https://review.openstack.org/43872 | 11:19 |
| *** SergeyLukjanov has joined #openstack-metering | 11:20 | |
| *** zbitter is now known as zaneb | 11:20 | |
| EmilienM | blinky_ghost: Ceilometer provides an API to get measurements. That's it | 11:23 |
| *** mmcardle has quit IRC | 11:36 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Implement testing of migrations https://review.openstack.org/44809 | 11:48 |
| *** ruhe has quit IRC | 12:01 | |
| *** ruhe has joined #openstack-metering | 12:04 | |
| *** ruhe has quit IRC | 12:06 | |
| *** shang has quit IRC | 12:17 | |
| *** ruhe has joined #openstack-metering | 12:17 | |
| *** thomasm has joined #openstack-metering | 12:18 | |
| thomasm | Hey all! | 12:21 |
| *** mmcardle has joined #openstack-metering | 12:25 | |
| *** gordc has joined #openstack-metering | 12:27 | |
| *** sandywalsh has quit IRC | 12:32 | |
| openstackgerrit | Julien Danjou proposed a change to openstack/ceilometer: Include auth_token middleware in sample config https://review.openstack.org/45886 | 12:38 |
| *** sandywalsh has joined #openstack-metering | 12:44 | |
| *** bpokorny has joined #openstack-metering | 12:48 | |
| *** shang has joined #openstack-metering | 12:53 | |
| *** mmcardle has quit IRC | 12:54 | |
| *** mmcardle has joined #openstack-metering | 12:59 | |
| *** ruhe has quit IRC | 13:01 | |
| *** ruhe has joined #openstack-metering | 13:03 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Implement testing of migrations https://review.openstack.org/44809 | 13:04 |
| *** mmcardle has quit IRC | 13:06 | |
| *** anteaya_ has joined #openstack-metering | 13:06 | |
| *** eglynn has quit IRC | 13:08 | |
| *** mmcardle has joined #openstack-metering | 13:10 | |
| *** anteaya_ is now known as anteaya | 13:13 | |
| *** shaneduan has quit IRC | 13:29 | |
| *** shaneduan has joined #openstack-metering | 13:29 | |
| *** shaneduan is now known as Guest77414 | 13:29 | |
| *** ruhe has quit IRC | 13:30 | |
| *** ruhe has joined #openstack-metering | 13:33 | |
| *** terriyu has joined #openstack-metering | 13:35 | |
| *** nosnos has quit IRC | 13:38 | |
| *** eglynn has joined #openstack-metering | 13:42 | |
| *** Ruetobas has joined #openstack-metering | 13:45 | |
| *** Ruetobas has quit IRC | 13:50 | |
| *** Ruetobas has joined #openstack-metering | 13:58 | |
| *** ruhe has quit IRC | 14:17 | |
| *** SergeyLukjanov has quit IRC | 14:17 | |
| *** mmcardle has quit IRC | 14:20 | |
| *** Ruetobas has quit IRC | 14:21 | |
| *** ruhe has joined #openstack-metering | 14:22 | |
| *** mmcardle has joined #openstack-metering | 14:22 | |
| *** Ruetobas has joined #openstack-metering | 14:23 | |
| *** ruhe has quit IRC | 14:26 | |
| *** SergeyLukjanov has joined #openstack-metering | 14:28 | |
| *** ruhe has joined #openstack-metering | 14:32 | |
| *** ruhe has quit IRC | 14:34 | |
| *** sandywalsh has quit IRC | 14:42 | |
| *** fnaval_ has joined #openstack-metering | 14:53 | |
| *** lexx has quit IRC | 14:57 | |
| *** sandywalsh has joined #openstack-metering | 14:59 | |
| *** SergeyLukjanov has quit IRC | 15:01 | |
| jd__ | dhellmann: did you send a request to update requirements wrt six/wsme? | 15:01 |
| jd__ | it seems not | 15:02 |
| jd__ | doing it then! | 15:02 |
| openstackgerrit | gordon chung proposed a change to openstack/ceilometer: validate counter_type when posting samples https://review.openstack.org/45344 | 15:03 |
| *** mmcardle has quit IRC | 15:03 | |
| dhellmann | jd__: I'm still working on making the wsme release work right | 15:03 |
| *** SergeyLukjanov has joined #openstack-metering | 15:03 | |
| dhellmann | we need to add some stuff to our tox config | 15:03 |
| jd__ | dhellmann: what does that mean? | 15:03 |
| dhellmann | I'm doing that now | 15:03 |
| dhellmann | http://logs.openstack.org/b0/b02f5f2375902448a6e3b379692a9110de2856a2/pre-release/wsme-tarball/41da501/console.html | 15:03 |
| jd__ | ok, it fails when you pushed the tag as I understand it | 15:04 |
| dhellmann | yeah | 15:04 |
| jd__ | that was not expected :] | 15:04 |
| dhellmann | so, new changeset on the way with the updated tox.ini and then a new tag | 15:04 |
| *** ruhe has joined #openstack-metering | 15:04 | |
| dhellmann | I'm looking at how to test the change locally, so give me a few minutes | 15:04 |
| jd__ | dhellmann: can't you push -f the tag? | 15:04 |
| dhellmann | that doesn't retrigger the job, iiuc | 15:05 |
| *** sandywalsh has quit IRC | 15:05 | |
| jd__ | dhellmann: ok | 15:05 |
| jd__ | well anyway a recheck no bug will be enough of the requirement patchset I've sent | 15:05 |
| dhellmann | jd__: https://review.openstack.org/46082 | 15:09 |
| openstackgerrit | gordon chung proposed a change to openstack/ceilometer: add tests for _query_to_kwargs func https://review.openstack.org/43796 | 15:09 |
| jd__ | dhellmann: LGTM | 15:10 |
| * dhellmann waits for jenkins to run the tests | 15:10 | |
| *** sandywalsh has joined #openstack-metering | 15:21 | |
| *** mmcardle has joined #openstack-metering | 15:28 | |
| *** mmcardle has quit IRC | 15:42 | |
| sandywalsh | eglynn, is there a reason https://github.com/openstack/ceilometer/blob/master/ceilometer/alarm/rpc.py#L40 lives in a different directory than other alarm notifiers and has a different signature? (is it unused?) | 15:45 |
| sandywalsh | dhellmann, ping? | 15:45 |
| eglynn | sandywalsh: that's simply the glue for the theshold evaluator to call out to the alarm notifier over RPC | 15:46 |
| eglynn | sandywalsh: (then the alarm notifier handles the actual notification of the end-consumer) | 15:46 |
| sandywalsh | eglynn, right, but isn't it the same as the other alarm notifiers? | 15:47 |
| sandywalsh | oh, I think I see the distinction | 15:47 |
| sandywalsh | there are two notifiers | 15:47 |
| eglynn | sandywalsh: not really | 15:47 |
| eglynn | sandywalsh: the RPC is an intermediate stage | 15:47 |
| eglynn | sandywalsh: the notifiers then handle the call out to the final consumer | 15:48 |
| sandywalsh | k, I see the difference there now. Thanks | 15:48 |
| sandywalsh | I saw the rpc call, but assumed there was only one style of notification (an email could be done directly if desired) | 15:48 |
| sandywalsh | but I see now that they all go through rpc and then relayed | 15:49 |
| thomasm | brb | 15:51 |
| *** thomasm has quit IRC | 15:52 | |
| *** eglynn is now known as eglynn-on-a-call | 15:53 | |
| *** thomasm has joined #openstack-metering | 15:54 | |
| nijaba | hello. Still waiting for another +2 on https://review.openstack.org/#/c/44774/.... anyone? | 15:55 |
| *** giroro_ has joined #openstack-metering | 16:03 | |
| *** Ruetobas has quit IRC | 16:03 | |
| *** giroro_ has quit IRC | 16:08 | |
| *** mmcardle has joined #openstack-metering | 16:11 | |
| *** fnaval_ has quit IRC | 16:12 | |
| *** Ruetobas has joined #openstack-metering | 16:14 | |
| *** fnaval_ has joined #openstack-metering | 16:15 | |
| *** fnaval__ has joined #openstack-metering | 16:19 | |
| *** fnaval_ has quit IRC | 16:21 | |
| *** mmcardle has quit IRC | 16:21 | |
| *** fnaval__ has quit IRC | 16:24 | |
| *** boris-42 has quit IRC | 16:29 | |
| *** Guest77414 is now known as shaneduan | 16:35 | |
| *** nati_ueno has joined #openstack-metering | 16:57 | |
| *** krtaylor has quit IRC | 17:03 | |
| openstackgerrit | Thomas Maddox proposed a change to openstack/ceilometer: Fix for timestamp precision in SQLAlchemy https://review.openstack.org/46100 | 17:04 |
| *** mmcardle has joined #openstack-metering | 17:04 | |
| *** mmcardle has quit IRC | 17:05 | |
| openstackgerrit | Thomas Maddox proposed a change to openstack/ceilometer: Fix for timestamp precision in SQLAlchemy https://review.openstack.org/46100 | 17:06 |
| openstackgerrit | Thomas Maddox proposed a change to openstack/ceilometer: Fix for timestamp precision in SQLAlchemy https://review.openstack.org/46100 | 17:07 |
| *** SergeyLukjanov has quit IRC | 17:09 | |
| openstackgerrit | Thomas Maddox proposed a change to openstack/ceilometer: Fix for timestamp precision in SQLAlchemy https://review.openstack.org/46100 | 17:10 |
| openstackgerrit | Thomas Maddox proposed a change to openstack/ceilometer: Fix for timestamp precision in SQLAlchemy https://review.openstack.org/46100 | 17:10 |
| thomasm | Pardon the spam, heh. | 17:11 |
| *** ruhe has quit IRC | 17:14 | |
| thomasm | wow Jenkins hated that | 17:23 |
| *** ruhe has joined #openstack-metering | 17:25 | |
| *** krtaylor has joined #openstack-metering | 17:44 | |
| *** dhellmann is now known as dhellmann_ | 17:47 | |
| *** boris-42 has joined #openstack-metering | 18:00 | |
| *** ruhe has quit IRC | 18:05 | |
| *** fnaval_ has joined #openstack-metering | 18:07 | |
| *** boris-42 has quit IRC | 18:15 | |
| *** boris-42 has joined #openstack-metering | 18:17 | |
| *** lnxnut has joined #openstack-metering | 18:20 | |
| *** SergeyLukjanov has joined #openstack-metering | 18:30 | |
| *** blinky_ghost has quit IRC | 18:34 | |
| *** yanghy has quit IRC | 18:41 | |
| dhellmann_ | sandywalsh: much delayed pong | 18:49 |
| sandywalsh | heh | 18:59 |
| dhellmann_ | story of my life lately | 19:00 |
| *** dhellmann_ is now known as dhellmann | 19:00 | |
| sandywalsh | np, I was wondering about the per-meter support in the publisher | 19:00 |
| sandywalsh | is the intention that we can have a topic per meter in the publisher? | 19:00 |
| dhellmann | nijaba: I see some other comments on that ticket. I can +2 to stop the bikeshedding if you want to keep working on it in another changeset? | 19:01 |
| dhellmann | sandywalsh: so that messages from different publishers can be sent to different amqp topics? | 19:01 |
| *** boris-42 has quit IRC | 19:01 | |
| sandywalsh | yeah, that's what I assumed it was for? Is that correct? | 19:01 |
| dhellmann | does common.rpc support that? | 19:02 |
| dhellmann | I think we want it to, but we might have to wait for oslo.messaging | 19:02 |
| sandywalsh | well, rpc can do it (you can specify the topic name, that's how the rpc layer does the ephemeral queues for the return values) | 19:02 |
| dhellmann | are we using rpc there, or notifications? | 19:03 |
| sandywalsh | but I'm wondering if we'd need 1 collector per topic on the receiving end | 19:03 |
| sandywalsh | it's rpc (a cast()) | 19:03 |
| dhellmann | it seems like notify must support that, too | 19:03 |
| sandywalsh | yeah, notify can do it too | 19:03 |
| dhellmann | oh, right, that's one of the things I wanted to change in icehouse | 19:03 |
| dhellmann | it might just be that we hadn't thought of that case yet | 19:03 |
| sandywalsh | I think we need to change all of our uses of cast() to notify() anyway. cast() isn't very reliable and there's no retry support | 19:03 |
| dhellmann | at first we were focusing on publishing using entirely different protocols | 19:04 |
| dhellmann | sandywalsh: yep | 19:04 |
| sandywalsh | the problem I can see is you'd need a ton of collectors if it's topic per meter | 19:04 |
| sandywalsh | (since there's a lot of meters) | 19:04 |
| dhellmann | well, I don't think we'd have them all be different by default, but we may want to allow different pipelines to output to different topics for other purposes | 19:05 |
| sandywalsh | the collector can take a list of topics to listen on I think as well ... and you can wildcard topics too. | 19:05 |
| sandywalsh | do you happen to recall who was driving that feature? I'd like to know how they're deploying it? | 19:06 |
| dhellmann | yeah | 19:06 |
| dhellmann | the multipublisher stuff? | 19:06 |
| dhellmann | that was added, in part, because of your insistence that we stop using rpc for all of the monitoring messages :-) | 19:06 |
| sandywalsh | I assumed "multipublisher" was more than one publisher ... vs topic-per-meter | 19:06 |
| dhellmann | I think jd__ wrote the UDP publisher | 19:06 |
| dhellmann | I'm not aware of a topic-per-meter feature per se | 19:07 |
| sandywalsh | lemme get a link, sec | 19:07 |
| sandywalsh | I think we may be talking about different things | 19:07 |
| dhellmann | I had some blueprints to clean up how meters and plugins were related | 19:07 |
| dhellmann | because I was trying to clean up the configuration loop that builds the pipelines | 19:07 |
| sandywalsh | https://github.com/openstack/ceilometer/blob/master/ceilometer/publisher/rpc.py#L116 | 19:08 |
| dhellmann | oh | 19:08 |
| dhellmann | I have no memory of that change. git blame? | 19:09 |
| sandywalsh | topic.meter https://github.com/openstack/ceilometer/blob/master/ceilometer/publisher/rpc.py#L173 | 19:09 |
| sandywalsh | yeah, that was the next step | 19:09 |
| sandywalsh | :) | 19:09 |
| sandywalsh | jd__ | 19:09 |
| sandywalsh | k, I'll ask him tomorrow when he's on ... just wondering if you knew off the top of your head what the rationale was | 19:10 |
| dhellmann | maybe someone thought it worked like notifications? | 19:10 |
| dhellmann | or maybe they are consuming the rpc messages as though they are notifications | 19:11 |
| sandywalsh | perhaps, or they were thinking it was a better way to scale out? | 19:11 |
| dhellmann | but that's just a guess | 19:11 |
| sandywalsh | not sure, more curious than anything. | 19:11 |
| dhellmann | by running lots of collector instances? maybe. I think we're already set up to pool, though, so having a single topic/exchange/queue should work better for scaling | 19:11 |
| dhellmann | yeah, now I am, too | 19:12 |
| sandywalsh | :) | 19:12 |
| sandywalsh | I almost have that alarm doc written up but I went down a rabbit hole documenting the current architecture. Might be handy for the general docs? | 19:12 |
| sandywalsh | and I stumbled across that and got confused :) | 19:12 |
| Guest21363 | dhellmann: do you have a minute to elaborate on the "disable event collection" question I raised? I was working under the understanding that pipeline config only impacted samples, i.e. not the notifications published by the dispatcher. | 19:15 |
| Guest21363 | argh, just a sec, let me authenticate (nealph) | 19:16 |
| *** SergeyLukjanov has quit IRC | 19:16 | |
| *** boris-42 has joined #openstack-metering | 19:17 | |
| sandywalsh | Guest21363/nealph, I put that in. The idea was that we only store events if that flag is True. Off by default. Less risk for new code. | 19:17 |
| sandywalsh | the samples should still work fine | 19:17 |
| sandywalsh | unless something busted during the dispatcher changes | 19:17 |
| *** nealph has joined #openstack-metering | 19:18 | |
| *** lnxnut has quit IRC | 19:19 | |
| nealph | sandywalsh: different topic, I think. I'm dovetailing the mailing list question I raised... | 19:20 |
| dhellmann | nealph: what is it you're trying to do? | 19:20 |
| *** Guest21363 has quit IRC | 19:21 | |
| nealph | which is, how to disable the collection of notifications from some services, i.e. limit to just Glance | 19:21 |
| dhellmann | nealph: the pipeline collects data for meters specified in that list, with * meaning "all" -- remove the * and include just the meters you want | 19:22 |
| *** SergeyLukjanov has joined #openstack-metering | 19:22 | |
| dhellmann | you don't have to know which notifications go into collecting the meter, you specify the meter by its short name | 19:22 |
| dhellmann | what's the easiest way to get a list of the reviews for items with FFE? I'd like to do some reviews, but we have a bunch of stuff in the queue and sorting through it all feels like a waste of time. | 19:23 |
| nealph | The issue here is that I'm consuming the dispatcher output, not the publisher output. Still effective? | 19:23 |
| sandywalsh | hmm, yeah, the dispatch will record all events ... there's no filter | 19:24 |
| dhellmann | yeah, the pipeline happens inside the dispatcher, doesn't it? | 19:24 |
| sandywalsh | might be an interesting addition to dragondm's trigger pipeline stuff ... but his stuff kicks in after the db has written the event. | 19:24 |
| dhellmann | nealph: are you writing your own plugin? | 19:25 |
| dhellmann | sandywalsh: the notifications go through a pipeline, which sends rpc messages, which go to the dispatcher, which go to the storage plugin | 19:27 |
| dhellmann | so if the pipeline only listens to the desired notifications, then the plugin on the other end of that chain should only get the desired meter data | 19:27 |
| dhellmann | now, that's only for samples/meter data, not for the events | 19:27 |
| sandywalsh | dhellmann, for samples, yes, but for events they store all events. nealph are you concerned about sample storage or event storage? | 19:28 |
| nealph | samples for now, as I believe that's all the glance pollster generates. | 19:29 |
| dhellmann | we get CRUD events for images from glance, too, I think | 19:29 |
| sandywalsh | gotcha ... glance can produce usage notifications though (outside of the pollster) | 19:29 |
| dhellmann | there's a notification plugin to subscribe to them, for sure | 19:29 |
| dhellmann | the only thing we poll for is image (existance) and image size | 19:30 |
| sandywalsh | https://github.com/openstack/ceilometer/blob/master/ceilometer/image/notifications.py | 19:31 |
| sandywalsh | assuming glance is configured to output notifications | 19:32 |
| nealph | well, when we're talking CRUD events on the bus from Glance, aren't those "events" (i.e. not "samples") | 19:32 |
| dhellmann | terminology gets confusing | 19:33 |
| nealph | or am I confusing the definition of "sample" (again :)) | 19:33 |
| nealph | dhellmann: extremely :) | 19:33 |
| dhellmann | we process 2 types of messages | 19:34 |
| dhellmann | notifications sent from outside services like glance | 19:34 |
| dhellmann | and rpc messages containing a single sample value coming from some part of ceilometer | 19:34 |
| dhellmann | notifications are "events" in the sense that they tell you something happened | 19:35 |
| dhellmann | the internal messages are "samples" in the sense that they tell you about one data point | 19:35 |
| dhellmann | which ones do you care about? | 19:35 |
| nealph | dhellmann: I need to know exists/delete/create, plus size, type, and all the meta data. | 19:36 |
| dhellmann | ok, both sets of messages can potentially tell you that | 19:37 |
| *** ruhe has joined #openstack-metering | 19:37 | |
| dhellmann | the notification messages will have it all in one package, and the samples will have each thing split out with a separate meter and value (one per message) | 19:37 |
| dhellmann | or maybe several per message, but clearly structured, I forget if we optimized that to send more than one sample at a time via rpc | 19:37 |
| nealph | so, size being one sample (gauge?), | 19:38 |
| dhellmann | right | 19:38 |
| dhellmann | nealph: are you trying to intercept the messages and process them directly, or do you want to have ceilometer store them and then use the API? | 19:40 |
| nealph | great question: intercept and process directly. | 19:40 |
| *** ruhe has quit IRC | 19:41 | |
| nealph | hence my questions around the dispatcher. | 19:41 |
| dhellmann | ok, we can work out how to do that, can I ask why? | 19:41 |
| dhellmann | yeah, writing a dispatcher plugin may be the easiest way to do what you want | 19:41 |
| dhellmann | the notification events are turned into sample messages in the collector using a pipeline, so you can still configure that to only give you the messages you want | 19:42 |
| nealph | As for why...to support the post-ceilometer processing I need to be able to output to file. Not very elegant, but it's the first step. :) | 19:43 |
| dhellmann | nealph: ok | 19:44 |
| dhellmann | nealph: I'm looking at the subscription code to figure out if configuring the pipeline has any impact on the dispatcher stuff | 19:44 |
| dhellmann | give me a minute or two to figure that out | 19:44 |
| nealph | Sure...many thanks. | 19:45 |
| dhellmann | ok, I think I've got it | 19:45 |
| dhellmann | let me get some links for you to look at | 19:45 |
| dhellmann | https://git.openstack.org/cgit/openstack/ceilometer/tree/ceilometer/collector/service.py | 19:46 |
| dhellmann | line 190 is the RPC endpoint called when ceilometer sends a message to itself | 19:46 |
| dhellmann | line 195 is the method called when a notification comes in | 19:46 |
| dhellmann | line 170 is the method that subscribes to the notifications | 19:47 |
| *** boris-42 has quit IRC | 19:47 | |
| dhellmann | line 244 is the loop where the notifications are given to the dispatchers | 19:50 |
| dhellmann | sorry, no, wrong line | 19:50 |
| dhellmann | sigh, no, it's doing that and building the list of problem events as a side-effect, so that's the right line | 19:50 |
| dhellmann | ok, so, what that tells me is that the dispatcher is given every notification that we receive | 19:51 |
| dragondm | Just noticed the ping: Eventually we will have a pipeline for events as well as samples (working on that now) That would be able to filter out any events you don't want. | 19:52 |
| dhellmann | nealph: so if you want to filter some of them you'll have to do it yourself for now | 19:52 |
| nealph | dragondm: despite all your earlier descriptions I didn't catch that you were including filters on specific events (only on the traits in the events).... | 19:53 |
| dragondm | I'm talking about a different component :> | 19:53 |
| dragondm | The code that's out there now is just for notification -> event conversion. I'm also working on a pipeline for events, similar to our sample pipelines. | 19:54 |
| nealph | dragondm: well I'll be the first to sign up, it seems. :) Are you doing any extensions to the existing dispatchers (database, file)? | 19:56 |
| dragondm | there will be some changes, I think. Was talking to sandywalsh this morning abt that. Mostly for the 3rd component of this, which is triggers (ability to collect events for later batch processing when some criteria are met) | 19:58 |
| dragondm | Those will require interaction w/ the dispatchers. | 19:58 |
| nealph | dhellmann: dragondm: okay...thanks all. I'm going to go work out some extensions to the existing dispatcher to see if I can wrangle it to do what I need. | 20:07 |
| *** anteaya has quit IRC | 20:21 | |
| *** SergeyLukjanov has quit IRC | 20:35 | |
| *** eglynn-on-a-call is now known as eglynn | 20:57 | |
| *** changbl has joined #openstack-metering | 21:06 | |
| *** herndon has joined #openstack-metering | 21:08 | |
| *** krtaylor has quit IRC | 21:09 | |
| *** changbl has quit IRC | 21:22 | |
| openstackgerrit | Doug Hellmann proposed a change to openstack/ceilometer: Add docstrings to some methods https://review.openstack.org/46141 | 21:36 |
| *** thomasm has quit IRC | 21:37 | |
| *** anteaya has joined #openstack-metering | 21:39 | |
| openstackgerrit | gordon chung proposed a change to openstack/ceilometer: Pecan assuming meter names are extensions https://review.openstack.org/46143 | 21:51 |
| *** shang has quit IRC | 21:55 | |
| *** bpokorny has quit IRC | 22:04 | |
| nijaba | dhellmann: I'll work on adressing the other comments in a new patch. thanks for your +2 | 22:06 |
| openstackgerrit | A change was merged to openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/45175 | 22:08 |
| *** shardy is now known as shardy_afk | 22:17 | |
| *** krtaylor has joined #openstack-metering | 22:18 | |
| *** changbl has joined #openstack-metering | 22:38 | |
| *** sandywalsh has quit IRC | 22:44 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Update the high level architecture https://review.openstack.org/44774 | 22:53 |
| openstackgerrit | A change was merged to openstack/python-ceilometerclient: alarm: rename counter_name to meter_name https://review.openstack.org/45076 | 22:54 |
| *** sandywalsh has joined #openstack-metering | 22:56 | |
| *** shang has joined #openstack-metering | 22:59 | |
| *** dhellmann is now known as dhellmann_ | 23:02 | |
| *** dhellmann_ is now known as dhellmann | 23:02 | |
| *** zbitter has joined #openstack-metering | 23:11 | |
| nealph | sandywalsh: ping...quick question. | 23:11 |
| *** dhellmann is now known as dhellmann_ | 23:11 | |
| nealph | to the room then: what's the expected output of the pipeline publisher file writer? | 23:12 |
| nealph | I assumed it would be sample content, but that's not what I'm seeing.... | 23:12 |
| *** zaneb has quit IRC | 23:14 | |
| *** gordc has quit IRC | 23:22 | |
| *** herndon has quit IRC | 23:27 | |
| *** anteaya_ has joined #openstack-metering | 23:35 | |
| *** anteaya has quit IRC | 23:38 | |
| *** anteaya__ has joined #openstack-metering | 23:55 | |
| *** anteaya_ has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!