Tuesday, 2016-03-01

*** rbak has quit IRC00:00
*** thorst has quit IRC00:01
openstackgerritKairo Araujo proposed openstack/ceilometer: hacking: check for assertTrue(isinstance(a, b))  https://review.openstack.org/28443300:12
*** Qiming has quit IRC00:17
*** cdent has joined #openstack-telemetry00:22
*** nicodemus_ has quit IRC00:29
*** cdent has quit IRC00:30
*** boris-42 has quit IRC00:34
*** claudiub has quit IRC00:42
*** thorst_ has quit IRC01:00
*** thorst has joined #openstack-telemetry01:01
*** thorst has quit IRC01:09
*** Qiming has joined #openstack-telemetry01:13
*** cheneydc has joined #openstack-telemetry01:15
openstackgerritliusheng proposed openstack/aodh: Add composite alarm usage description  https://review.openstack.org/28012101:15
*** thorst has joined #openstack-telemetry01:45
*** thorst has quit IRC01:53
*** thorst has joined #openstack-telemetry02:51
*** thorst has quit IRC02:58
*** achatterjee has joined #openstack-telemetry03:18
*** dave-mccowan has quit IRC03:45
*** links has joined #openstack-telemetry03:49
*** thorst has joined #openstack-telemetry03:56
*** chlong_ has quit IRC04:03
*** thorst has quit IRC04:03
*** yprokule has joined #openstack-telemetry05:00
*** thorst has joined #openstack-telemetry05:01
openstackgerritMerged openstack/ceilometer: Fix a minor missing parameter issue  https://review.openstack.org/28560605:04
*** thorst has quit IRC05:08
openstackgerritRabi Mishra proposed openstack/ceilometer: Change default policy to allow create_samples  https://review.openstack.org/28638505:14
ramishrahi,  anyone around to look at ^^^, it seems changes to default policy has broken heat tests.05:20
*** achatterjee has quit IRC05:35
*** thorst has joined #openstack-telemetry06:06
*** rcernin has joined #openstack-telemetry06:09
*** thorst has quit IRC06:13
openstackgerritRabi Mishra proposed openstack/ceilometer: Change default policy to allow create_samples  https://review.openstack.org/28638507:01
*** thorst has joined #openstack-telemetry07:11
tobascoif i'm running a ceilometer with a mongodb backend and has database.metering_time_to_live configured to some ttl value, do I have to call the ceilometer-expirer or will mongodb handle all expiring of data?07:11
*** jwcroppe has quit IRC07:13
silehttobasco, mongodb don't need it07:14
*** jwcroppe has joined #openstack-telemetry07:14
tobascosileht: thanks07:14
*** thorst has quit IRC07:18
openstackgerritZi Lian Ji proposed openstack/ceilometer: Enable the Load Balancer v2 events  https://review.openstack.org/28177707:32
*** belmoreira has joined #openstack-telemetry08:01
*** jwcroppe_ has joined #openstack-telemetry08:04
openstackgerritZi Lian Ji proposed openstack/ceilometer: Add the meter example file 'lbaas-v2-meter-definitions.yaml'  https://review.openstack.org/28643108:06
*** jwcroppe has quit IRC08:07
*** pcaruana has joined #openstack-telemetry08:09
*** thorst has joined #openstack-telemetry08:16
*** liamji has joined #openstack-telemetry08:17
*** thorst has quit IRC08:23
openstackgerritZi Lian Ji proposed openstack/ceilometer: Ceilometer: fix MeterDefinitionException for py3  https://review.openstack.org/28519608:38
*** _nadya_ has joined #openstack-telemetry08:39
*** AJaeger has joined #openstack-telemetry08:44
AJaegercould a stable maintainer import translations for liberty, please? This should be the last import for liberty, we're stopping translations now. https://review.openstack.org/#/c/280514/ jd__, eglynn, please help08:45
*** openstackgerrit has quit IRC08:48
*** openstackgerrit has joined #openstack-telemetry08:49
*** ildikov has quit IRC09:08
*** idegtiarov_ has joined #openstack-telemetry09:19
*** thorst has joined #openstack-telemetry09:22
*** thorst has quit IRC09:28
*** yassine__ has joined #openstack-telemetry09:32
*** liamji has quit IRC09:33
*** ildikov has joined #openstack-telemetry09:35
openstackgerritIgor Degtiarov proposed openstack/ceilometer: timedelta plugin for meter definition process  https://review.openstack.org/28646909:41
permalachello guys , just installed kilo on centos 7   by the book, and ceilometer shows information on the horizon dashboard but reports 'The service catalog is empty' from command line 'ceilometer meter-list'09:52
permalacI've sourced the admin-rc env vars and so, but I do not know how to debug this. could you help ?09:52
*** cheneydc has quit IRC10:01
*** ildikov has quit IRC10:26
*** thorst has joined #openstack-telemetry10:26
*** Qiming has quit IRC10:30
*** therve has joined #openstack-telemetry10:32
*** thorst has quit IRC10:33
-openstackstatus- NOTICE: Gerrit is going to be restarted due to poor performance10:38
*** ChanServ changes topic to "Gerrit is going to be restarted due to poor performance"10:38
*** therve has left #openstack-telemetry10:38
*** ildikov has joined #openstack-telemetry10:40
*** ChanServ changes topic to "OpenStack Telemetry | http://wiki.openstack.org/Telemetry"10:44
-openstackstatus- NOTICE: gerrit finished restartign10:44
*** lalolmalo has quit IRC10:58
*** chlong_ has joined #openstack-telemetry11:01
*** cdent has joined #openstack-telemetry11:01
*** thorst has joined #openstack-telemetry11:31
*** ildikov has quit IRC11:36
*** Qiming has joined #openstack-telemetry11:36
*** thorst has quit IRC11:38
permalacguys, do you use mongo as backend ?12:11
permalacWhat howto can I follow to have it in HA , with just 3 servers .  ?12:12
openstackgerritMerged openstack/aodh: Remove unused pngmath Sphinx extension  https://review.openstack.org/28620012:16
*** dave-mccowan has joined #openstack-telemetry12:27
*** ildikov has joined #openstack-telemetry12:38
idegtiarovpermalac, hi! you could deploy mongo in replica mode12:40
*** gordc has joined #openstack-telemetry12:40
permalacidegtiarov, is what I'm reading now, still don't understand how the clients know where to write when the primary changes in case of failure.12:41
*** thorst has joined #openstack-telemetry12:46
*** shardy has quit IRC12:47
*** thorst_ has joined #openstack-telemetry12:47
*** jaypipes has joined #openstack-telemetry12:48
*** thorst has quit IRC12:51
idegtiarovpermalac, it is pymongo's headache12:53
permalacso, it just works, I give all the mongo server hostnames to the ceilometer.conf [database] section and it will work by itself?12:54
*** links has quit IRC13:07
*** julim has joined #openstack-telemetry13:10
idegtiarovpermalac, yep it will13:10
*** pradk has joined #openstack-telemetry13:14
*** nicodemus_ has joined #openstack-telemetry13:34
nicodemus_hello gordc13:40
gordcnicodemus_: hey, whatsup?13:43
nicodemus_gordc, not much... here to help13:44
gordccool cool. i won't lie, i just finished breakfast so i haven't done anything yet.13:45
gordcgoing to start with this ubuntu machine.13:45
nicodemus_gordc: haha don't worry13:45
nicodemus_I don't intend to be a pain in the lower back13:46
gordclol13:47
gordci might have questions for you later... or probably tomorrow13:47
*** fwdit has joined #openstack-telemetry13:47
*** shardy has joined #openstack-telemetry13:50
permalacidegtiarov, thanks13:53
*** eglynn has quit IRC14:11
openstackgerritMerged openstack/gnocchi: fix migration from v1.3 to v2.0  https://review.openstack.org/28452814:11
openstackgerritMerged openstack/ceilometer: Enable the Load Balancer v2 events  https://review.openstack.org/28177714:20
*** rickyrem has joined #openstack-telemetry14:22
*** rickyrem1 has joined #openstack-telemetry14:26
*** rickyrem has quit IRC14:27
openstackgerritMerged openstack/ceilometer: Change default policy to allow create_samples  https://review.openstack.org/28638514:28
*** jroll has quit IRC14:29
*** jroll has joined #openstack-telemetry14:30
*** jroll has quit IRC14:30
*** jroll has joined #openstack-telemetry14:30
idegtiarovpermalac, np14:30
*** shardy has quit IRC14:33
*** shardy has joined #openstack-telemetry14:35
*** rbak has joined #openstack-telemetry14:46
*** yprokule has quit IRC14:50
openstackgerritIgor Degtiarov proposed openstack/ceilometer: timedelta plugin for meter definition process  https://review.openstack.org/28646914:51
*** yprokule has joined #openstack-telemetry14:56
idegtiarovgordc, hi!14:58
gordcidegtiarov: hi14:58
gordcwhatsup?14:59
idegtiarovgordc, could you please take a look on patch https://review.openstack.org/286469 the same result as we going achive with event transformers but without events without transformations without any cache14:59
*** shardy has quit IRC15:00
idegtiarovgordc,  and rather small15:00
idegtiarovgordc, the question is could we have it in M?15:01
gordcidegtiarov: seems simple enough15:03
gordcdoes it actually work? i wasn't aware the plugin stuff worked for meters.15:04
idegtiarovgordc, yes it works15:05
*** liamji has joined #openstack-telemetry15:06
gordci'm fine with it. as long as this doesn't grow in scope, i'm ok with adding it as FFE15:06
idegtiarovgordc, cool thanks15:06
gordcyou need to expand tests though. i'll add a few comments15:07
idegtiarovgordc,  will do15:07
idegtiarovshould I move code to new meter_plugin module15:07
idegtiarovor it ok that it inherits event.trait15:08
gordcidegtiarov: i think it's ok to leave it. maybe add a TODO to refactor it better.15:18
idegtiarovkbyrne, thanks15:18
idegtiarovgordc, k thanks15:18
gordcnp15:18
*** rickyrem1 has quit IRC15:19
gordcsileht: really weird the gnocchi dispatcher loses connection in gate.15:20
gordcdon't know why i don't see it locally15:20
*** AJaeger has left #openstack-telemetry15:21
*** pradk has quit IRC15:22
gordcityaptin_: do you have time to add a test for https://review.openstack.org/#/c/285431/?15:22
*** pradk has joined #openstack-telemetry15:22
ityaptin_gordc: yep, it's in progress now15:22
*** jdandrea_ has joined #openstack-telemetry15:23
*** fwdit has quit IRC15:23
permalacI have a mongodb primary RS and two other servers as slave,  how do I set this up on the ceilometer.conf ?15:23
gordcityaptin_: awesome. it seems to look good to me.15:24
*** jdandrea has quit IRC15:26
openstackgerritMerged openstack/aodh: Add composite alarm usage description  https://review.openstack.org/28012115:32
*** rickyrem has joined #openstack-telemetry15:39
*** liamji has quit IRC15:42
*** ryanpetrello has quit IRC15:43
*** ryanpetrello has joined #openstack-telemetry15:44
*** jaypipes has quit IRC15:50
*** shardy has joined #openstack-telemetry15:51
*** KrishR has joined #openstack-telemetry15:55
permalaccan someone explain me how to check if ceilometer is receiving new data?15:59
permalaccan I trigger a data-collector job ?15:59
*** yprokule has quit IRC16:01
gordcpermalac: what data are you looking for?16:08
permalacI've killed a mongodb node, and I don't see any new data on the horizon ceilometer plugin , so I would like understand how the new data is processed to find where I missplaced the configuration16:09
*** chlong_ has quit IRC16:10
gordci should tell you first that mongodb has potential scaling issues... and also the horizon plugin is being disabled as it's not actually functional16:11
permalacouch!16:11
gordcthat said, you should really consider what metrics you want. http://docs.openstack.org/admin-guide-cloud/telemetry-measurements.html16:11
gordcsome metrics are gathered via polling. others are built from notifications from the services.16:12
permalacI'm in Kilo but we intend to standarize openstack as company platform, so I'm really interested on how to show the measurements if not with horizon.16:12
*** jaypipes has joined #openstack-telemetry16:13
gordcif it's kilo, you can continue with horizon+ceilometer. how functionaly it is, will be debatable.16:14
gordcalso, kilo will be eol in the coming weeks16:14
permalacit's not even nice, I agree with that. But if we are just starting may be we should already move to whatever will be in the future.16:14
gordcpermalac: that makes sense. i think you can get support for mongodb but i believe people have started to look at alternatives.16:15
gordcwe have Gnocchi (gnocchi.xyz) which is one alternative.16:16
gordcyou can also look at other tsdb but that you'd have to build yourself.16:16
permalacI found what I did wrong, I changed the mongodb connection string but I did not restart collector . now it works.16:16
permalactsdb ? whats that?16:16
gordctime series db16:17
*** Qiming has quit IRC16:17
permalacand what about ELK or similar tools , do someone use it ?16:17
gordcsomething like prometheus or opentsdb16:17
gordcwe support elasticsearch for our events (open-text, log based data)16:18
gordcit apparently can support metric data but that could just be a marketing point16:18
gordci would just start with a db you're familiar with.16:19
gordcceilometer doesn't make any effort to scale dbs. the data is just passed to whatever db you select with the assumption you know how to operate that db16:20
permalacbut I don't mind the db, I would like a nice way to extrat information from all the data.16:20
permalac*extract16:20
gordcwhat's 'all the data'?16:20
permalacthe metric data,16:21
permalacwe will store 6 months of that for report pourposes, probably on a secondary database instance with some delay , but I would like to have a nice way to see the statistics and reports as for billing pourposes16:22
permalacbilling based on usage will be the end of the road.16:22
gordcpermalac: if you want all the data (including it's metadata) you will run into scaling problems. the standard dbs ceilometer provides store everything, so it will be very verbose and performance may be a concern.16:23
gordcif you want performant results over a longer time period (such as 6 months), you'll want to look at a tsdb which will aggregate and optimise the data so it doesn't' grow uncontrollably16:24
gordcs/standard/legacy/16:25
*** belmoreira has quit IRC16:28
*** _nadya_ has quit IRC16:28
*** _nadya_ has joined #openstack-telemetry16:29
*** _nadya_ has quit IRC16:33
*** thorst_ is now known as thorst_afk16:47
*** rcernin has quit IRC16:55
*** pcaruana has quit IRC17:07
*** anteaya has joined #openstack-telemetry17:15
anteayahello telemetry17:16
anteayaI attended the ops meetup in manchester at which there was a scientific working group session17:16
anteayaI had agreed to bring some ceilometer feedback from the ops back to devs17:17
anteayahere is the link to the etherpad for the scientific working group session: https://etherpad.openstack.org/p/MAN-ops-Scientific-WG17:17
anteayaon line 116 there is a section for monitoring tools17:17
anteayathe overall feeling from the room is that ceilometer is slow17:18
anteayaregardless of the backend17:18
anteayato the point that some were indicating that they were using other solutions or looking to find other solutions17:18
anteayaI'll link to this channel log in my follow up to the operators list to share with operators that I brought this feedback to the developers17:19
anteayathanks for listening17:19
gordcanteaya: thanks for the feedback17:19
anteayagordc: would you like to respond in any way?17:19
gordcanteaya: i believe everyone is aware of the storage issue17:19
anteayaeither in channel or on the post I create to the ops list?17:20
anteayaokay17:20
anteayaare there any specs or blueprints up that I can point operators to?17:20
gordcanteaya: to be honest, it'd be better to get feedback here rather than a generic dump every yera17:20
anteayaokay17:20
gordcanteaya: we've been working on gnocchi.17:21
anteayaso how would you like me to reply to operators when I talk to them17:21
anteayawhat is gnocchi?17:21
gordcanteaya: i know others have their own drivers using hbase or opentsdb17:21
anteayais performance different with hbase or opentsdb?17:21
gordcanteaya: 'please ask questions in the list or irc' that'd be a start17:22
anteayawell they don't know how17:22
anteayafor many of them this was the first time they had interacted with any developers17:22
gordcanteaya: also, we can't really do anything about mongodb. we don't own or develop mongodb so i'm not sure what action they'd like us to take on that17:23
anteayaand for the sceientific working group this was the first time they had met each other17:23
anteayawell it is less we want you to take action foo and more, we are using this and it isn't working17:23
gordcgnocchi is our attempt at building a time series driven storage, which does pre-aggregations17:24
anteayaso I said I appreciating hearing their experience and would do my best to communicate that17:24
anteayaokay17:24
anteayawhat stage is gnocchi at?17:24
gordcthis will effectively minimise the db size and response as it rollsup data and does all its computations before request rather than after request17:25
anteayagreat17:25
anteayawhat stage is it at?17:25
gordcanteaya: that's cool. thanks for communicating it. we like feedback, it's just very difficult to fix if we can't followup/dive-in to problems17:26
anteayaagreed17:26
gordcanteaya: gnocchi just released 2.017:26
gordcwe have a few people testing. thanks to them.17:26
anteayawho is testing it?17:26
gordcbut yeah, we just need testers.17:26
gordcnicodemus_: cloudkitty folks for one.17:27
gordcit's integrated in gate with heat/aodh.17:27
anteayado people need to be using heat and aodh to use gnocchi?17:27
gordcanteaya: no. it' just handles data storage (gnocchi.xyz)17:28
gordcaodh handles alarming, so it uses data stored in gnocchi17:28
gordcheat uses aodh for alarms17:28
anteayaI don't know what gnocchi.xyz is17:28
gordcceilometer uses gnocchi for storage17:28
anteayaso gnocchi is a backend?17:29
anteayadoes gnocchi replace mongodb for instance?17:29
gordcit's a website. xyz is one of those new fancy domain names17:29
anteayaoh17:29
gordcsame as http://docs.openstack.org/developer/gnocchi/17:29
gordcgnocchi would replace mongodb's metering capabilities yes.17:30
anteayayeah I don't understand why docs are in two places17:30
anteayabut I will defer that since I want to stay focused on operator feedback17:30
gordcit's the same docs. just published to multiple targets17:30
anteayayeah I don't understand that choice17:30
anteayaso is gnocchi faster than mongodb?17:30
anteayaso is the response to ceilometer is slow is to use gnocchi for storage?17:31
gordcdepends what you need.17:31
*** rickyrem has quit IRC17:31
anteayawell seems the scientific group would like speed17:32
gordcmentioned briefly in notes: http://docs.openstack.org/developer/ceilometer/architecture.html17:32
anteayawould they get speed out of using gnocchi?17:32
gordcspeed for querying responses?17:33
anteayayes that would count I believe17:33
*** KrishR has quit IRC17:33
*** rickyrem has joined #openstack-telemetry17:34
gordcanteaya: in theory yes, you have quicker queries since the statistic is already computed17:34
anteayagreat17:34
anteayahave any tests been able to confirm that yet?17:35
*** thorst_afk has quit IRC17:35
gordcanteaya: not one for one comparisons.17:35
anteayaokay thanks17:35
gordchttps://julien.danjou.info/blog/2015/gnocchi-benchmarks17:35
anteayathis is useful information17:36
*** thorst_afk has joined #openstack-telemetry17:36
gordcwe have benchmarks from a while back... but that's not on ceilometer data17:36
anteayaoh17:36
anteayaokay17:36
anteayathanks for the chat gordc17:36
*** thorst_afk is now known as thorst_17:36
anteayaI'll post the link to this log to operators and call this action item done17:36
gordcanteaya: np. i would recommend ops or anyone come visit us here17:37
gordcwe don't bite17:37
anteayagood to know17:37
gordcbut we are predominately asia/euro based17:37
anteayahopefully my post with get some folks to comment on the thread17:37
anteayaso maybe a conversation will get started17:37
anteayathanks gordc17:37
gordcanteaya: cool cool. thanks for providing feedback17:38
anteayahappy to help17:38
anteayagordc: the link to my ml post on -operators: http://lists.openstack.org/pipermail/openstack-operators/2016-March/009709.html17:43
gordcanteaya: great. thanks17:44
anteayawelcome17:47
*** KrishR has joined #openstack-telemetry18:03
*** tongli has joined #openstack-telemetry18:15
*** openstackgerrit has quit IRC18:18
*** openstackgerrit has joined #openstack-telemetry18:19
*** serverascode has quit IRC18:26
*** serverascode has joined #openstack-telemetry18:28
*** permalac has quit IRC18:31
*** permalac has joined #openstack-telemetry18:34
*** permalac has quit IRC18:34
*** idegtiarov_ has quit IRC18:42
*** wolsen has quit IRC18:50
*** afazekas has quit IRC18:52
*** john5223 has quit IRC18:52
*** openstack has joined #openstack-telemetry19:18
nicodemus_gordc: certainly19:18
nicodemus_gordc: please tell me your IP address, so that I set up the rule in the firewall19:19
*** openstackgerrit has joined #openstack-telemetry19:20
*** smurke_ has quit IRC19:25
*** sudorandom has quit IRC19:25
*** alextricity25 has quit IRC19:25
*** smurke_ has joined #openstack-telemetry19:31
*** sudorandom has joined #openstack-telemetry19:31
*** alextricity25 has joined #openstack-telemetry19:31
*** rickyrem1 has joined #openstack-telemetry19:31
*** rickyrem has quit IRC19:33
*** fc__ has quit IRC19:33
*** fc__ has joined #openstack-telemetry19:33
*** twm_ has joined #openstack-telemetry19:39
*** david-lyle has quit IRC19:59
*** julim has quit IRC20:07
*** julim has joined #openstack-telemetry20:08
*** david-lyle has joined #openstack-telemetry20:09
*** julim has quit IRC20:10
*** twm_ has quit IRC20:13
gordcnicodemus_: cool, i think i found where it's going nuts. just need to figure out what the right response is.20:30
gordcnicodemus_: thanks for lending me the machine.20:30
*** _nadya_ has quit IRC20:31
*** tongli has quit IRC20:35
nicodemus_gordc: no problem! anytime20:36
nicodemus_gordc: could you elaborate on where it's going nuts?20:37
*** alejandrito has joined #openstack-telemetry20:38
nicodemus_hi alejandrito20:38
nicodemus_hi alejandrito20:38
nicodemus_alejandrito: hi20:39
gordcnicodemus_: i need to play around with it a bit more but from what i can tell, round_timestamp is designed to handle single timestamp20:40
gordcnicodemus_: but when we do split https://github.com/openstack/gnocchi/blob/master/gnocchi/carbonara.py#L31820:40
gordcit's eventually sending the entire index of the series.20:40
gordci'm thinking we need to handle that in more than a single line.20:42
nicodemus_gordc: nice catch20:42
*** idegtiarov_ has joined #openstack-telemetry20:44
*** pcaruana has joined #openstack-telemetry20:44
*** julim has joined #openstack-telemetry20:55
*** cdent has quit IRC21:03
*** david-lyle has quit IRC21:08
*** david-lyle has joined #openstack-telemetry21:17
*** nicodemus_ has quit IRC21:25
*** KrishR has quit IRC21:38
*** r-mibu has quit IRC21:46
*** r-mibu has joined #openstack-telemetry21:47
*** idegtiarov_ has quit IRC21:49
*** pcaruana has quit IRC22:01
*** thorst_ has quit IRC22:09
*** shardy has quit IRC22:15
*** alejandrito has quit IRC22:18
openstackgerritMerged openstack/ceilometer: Add /usr/local/{sbin,bin} to rootwrap exec_dirs  https://review.openstack.org/28296222:29
*** david-lyle has quit IRC22:29
*** david-lyle has joined #openstack-telemetry23:00
*** yassine__ has quit IRC23:01
*** chlong_ has joined #openstack-telemetry23:13
*** Ephur has joined #openstack-telemetry23:16
*** rickyrem1 has quit IRC23:27
*** rbak has quit IRC23:39
*** david-lyle has quit IRC23:41
*** Qiming has joined #openstack-telemetry23:41
*** gordc has quit IRC23:53

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!