| *** xmltok has quit IRC | 00:56 | |
| openstackgerrit | A change was merged to openstack/python-ceilometerclient: Using common method 'bool_from_string' from oslo strutils https://review.openstack.org/68310 | 01:11 |
|---|---|---|
| openstackgerrit | A change was merged to openstack/ceilometer: Set the SQL Float precision https://review.openstack.org/63132 | 01:11 |
| openstackgerrit | Luis A. Garcia proposed a change to openstack/ceilometer: Sync gettextutils from Oslo https://review.openstack.org/66982 | 01:16 |
| openstackgerrit | Luis A. Garcia proposed a change to openstack/ceilometer: Re-enable lazy translation https://review.openstack.org/66983 | 01:16 |
| openstackgerrit | Lianhao Lu proposed a change to openstack/ceilometer: Drop foreign key contraints of alarm in sqlalchemy https://review.openstack.org/65018 | 01:24 |
| *** luisg has joined #openstack-ceilometer | 01:32 | |
| *** adriant has joined #openstack-ceilometer | 01:47 | |
| *** xianghui has joined #openstack-ceilometer | 02:19 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Exit expirer earlier if db-ttl is disabled https://review.openstack.org/67803 | 02:21 |
| openstackgerrit | A change was merged to openstack/ceilometer: Exclude weak datapoints from alarm threshold evaluation https://review.openstack.org/67161 | 02:26 |
| *** flwang has quit IRC | 02:28 | |
| *** flwang has joined #openstack-ceilometer | 02:43 | |
| *** yfujioka has joined #openstack-ceilometer | 02:57 | |
| *** thomasem has joined #openstack-ceilometer | 02:59 | |
| *** thomasem has quit IRC | 03:03 | |
| openstackgerrit | Rob Raymond proposed a change to openstack/ceilometer: Fix some typos in architecture doc https://review.openstack.org/68714 | 04:20 |
| *** adriant has quit IRC | 04:24 | |
| *** yfujioka has quit IRC | 05:02 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 05:22 | |
| openstackgerrit | Jenkins proposed a change to openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/62808 | 06:02 |
| openstackgerrit | Rob Raymond proposed a change to openstack/ceilometer: Correct spelling of logger for dispatcher.file https://review.openstack.org/68854 | 06:10 |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 06:29 | |
| *** nadya__ has joined #openstack-ceilometer | 06:49 | |
| *** nadya__ has quit IRC | 06:52 | |
| *** ildikov_ has joined #openstack-ceilometer | 06:54 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 07:22 | |
| *** lsmola_ has joined #openstack-ceilometer | 07:34 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_a | 07:39 | |
| *** SergeyLukjanov_a is now known as SergeyLukjanov_ | 07:40 | |
| *** vkodam has joined #openstack-ceilometer | 07:41 | |
| vkodam | hi jd___, dhellman | 07:42 |
| vkodam | what is the procedure to get the blueprint approved? | 07:42 |
| vkodam | I am referring to a blueprint which I have updated recently | 07:42 |
| vkodam | https://blueprints.launchpad.net/ceilometer/+spec/dynamic-meters | 07:42 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 07:49 | |
| *** s2r2_ has joined #openstack-ceilometer | 07:52 | |
| *** yassine has joined #openstack-ceilometer | 08:54 | |
| openstackgerrit | A change was merged to openstack/python-ceilometerclient: abbreviating --meter-name to -m in alarm commands https://review.openstack.org/66254 | 08:54 |
| openstackgerrit | A change was merged to openstack/python-ceilometerclient: Replace inheritance hierarchy with composition https://review.openstack.org/60857 | 08:54 |
| openstackgerrit | A change was merged to openstack/python-ceilometerclient: Add support for groupby in statistics for API v2 https://review.openstack.org/67433 | 08:55 |
| *** yassine has quit IRC | 09:03 | |
| *** yassine has joined #openstack-ceilometer | 09:03 | |
| *** yassine has quit IRC | 09:04 | |
| *** yassine has joined #openstack-ceilometer | 09:04 | |
| *** yassine has quit IRC | 09:05 | |
| openstackgerrit | Jia Dong proposed a change to openstack/ceilometer: Implement meter query by 'counter_volume' field https://review.openstack.org/67384 | 09:05 |
| *** yassine has joined #openstack-ceilometer | 09:06 | |
| *** yassine has quit IRC | 09:06 | |
| *** yassine has joined #openstack-ceilometer | 09:07 | |
| *** yassine has quit IRC | 09:07 | |
| *** yassine has joined #openstack-ceilometer | 09:08 | |
| *** yassine has quit IRC | 09:09 | |
| *** yassine has joined #openstack-ceilometer | 09:10 | |
| *** yassine has quit IRC | 09:12 | |
| *** yassine has joined #openstack-ceilometer | 09:12 | |
| *** yassine has quit IRC | 09:17 | |
| *** yassine has joined #openstack-ceilometer | 09:17 | |
| *** eglynn has joined #openstack-ceilometer | 09:21 | |
| openstackgerrit | Julien Danjou proposed a change to openstack/ceilometer: eventlet: stop monkey patching https://review.openstack.org/68145 | 09:24 |
| openstackgerrit | Julien Danjou proposed a change to openstack/ceilometer: Revert "Ensure we are not exhausting the sqlalchemy pool" https://review.openstack.org/68399 | 09:24 |
| eglynn | lsmola_: good morrow sir! | 09:25 |
| lsmola_ | eglynn: hello | 09:26 |
| jd__ | sigh, people | 09:26 |
| eglynn | lsmola_: ... so now that i-2 is done (such as it is) I hope to ring-fence off some time to look properly into the tuskar/horizon "distinct" requirement on the ceilo API that you spoke about before | 09:26 |
| eglynn | lsmola_: just wondering about how that lines up against other tuskar/horizon requirement(s) on ceilo | 09:27 |
| eglynn | (... priority-wise for the I cycle, that is) | 09:27 |
| eglynn | ... also is there an existing BP /me wonders? | 09:27 |
| lsmola_ | eglynn: for now, only the baremetal metrics are needed for Tuskar | 09:27 |
| eglynn | lsmola_: a-ha, k, llu's work in effect? | 09:28 |
| eglynn | lsmola_: ... so the disctinct thing is a pure horizon requirement? | 09:28 |
| ildikov_ | eglynn: we've had a plan to include this into our complex query for statistics bp | 09:28 |
| eglynn | lsmola_: ... i.e. the non-tuskar-y bit of horizon | 09:28 |
| lsmola_ | eglynn: those metrics shouldn't require any special querying, I should be able to aggregate them as i need now | 09:28 |
| eglynn | ildikov_: a-ha, I see, thanks! | 09:29 |
| lsmola_ | eglynn: yes, that is the Horizon's project and admin dashboard | 09:29 |
| eglynn | ildikov_: ... so my work here is done ;) | 09:29 |
| lsmola_ | eglynn: but we might need it in tuskar, if we propose more baremetal metrics | 09:29 |
| ildikov_ | eglynn: I'm not sure that there is the distinct, what we really need to use, but I have to rethink this as we the last discussion about it was a quite long time ago | 09:29 |
| eglynn | ildikov_: k | 09:30 |
| eglynn | lsmola_: ... "those metrics" == "samples for the instance meter and other meters that reflect resource lifecycle" ? | 09:30 |
| ildikov_ | eglynn: the bp is here: https://blueprints.launchpad.net/ceilometer/+spec/complex-query-support-for-statistics, targeted to i-3 | 09:30 |
| lsmola_ | ildikov_: yeah the distinct din't really suit the problem well, AFAIK | 09:31 |
| eglynn | ildikov_: /me thanks you ... and reads | 09:31 |
| ildikov_ | eglynn, lsomla_: I'm wondering, that what we can do with the periodic needs of the instance duration meter | 09:31 |
| eglynn | lsmola_: ... so are you saying that what ildikov_ is proposing will not fit well with the horizon requirement? | 09:31 |
| *** SergeyLukjanov is now known as SergeyLukjanov_a | 09:31 | |
| lsmola_ | eglynn: there are this 2 patches needed for baremetal metrics https://review.openstack.org/#/c/43074/ (and dep) | 09:32 |
| lsmola_ | eglynn: well | 09:32 |
| lsmola_ | eglynn: i am not exactly sure how to build such a query | 09:32 |
| *** SergeyLukjanov_a is now known as SergeyLukjanov_ | 09:32 | |
| lsmola_ | eglynn: so i don't know | 09:33 |
| lsmola_ | :-) | 09:33 |
| *** yassine has quit IRC | 09:33 | |
| *** yassine has joined #openstack-ceilometer | 09:33 | |
| ildikov_ | eglynn, lsmola_: I think the requirement should be reinvestigated, as according to my memories about this question, we could not find the query expression that would solve the issue perfectly | 09:33 |
| lsmola_ | ildikov_: yes | 09:33 |
| eglynn | so my original understand was that the basic requirement was along lines of ... | 09:33 |
| eglynn | "tell me how many *distinct* instances a tenant had running in each of these periods making up a longer-term duration" | 09:34 |
| eglynn | lsmola_: yea or nay? | 09:34 |
| lsmola_ | eglynn: yes i think | 09:35 |
| lsmola_ | eglynn: i had it somewhere in paste, let me dig | 09:35 |
| lsmola_ | eglynn: http://paste.openstack.org/show/54676/ | 09:36 |
| lsmola_ | eglynn: https://blueprints.launchpad.net/horizon/+spec/project-overview-page-ceilometer | 09:37 |
| eglynn | lsmola_: ok, that sounds my understanding too, thanking you ... | 09:37 |
| *** yassine has quit IRC | 09:37 | |
| *** yassine has joined #openstack-ceilometer | 09:38 | |
| nadya_ | hi all! | 09:39 |
| ildikov_ | lsoma_: if I remember right, we had a short discussion about this with my colleague, gibi and we found some issues regarding to use the instance meter, and how it would or would not solve this particular issue | 09:39 |
| *** yassine has quit IRC | 09:39 | |
| *** yassine has joined #openstack-ceilometer | 09:40 | |
| *** yassine has quit IRC | 09:42 | |
| *** yassine has joined #openstack-ceilometer | 09:42 | |
| ildikov_ | eglynn, lsmola_: do we have that meter, which shows that the instance is up or down (stopped or suspended)? | 09:47 |
| eglynn | ildikov_: well we have the instance meter, and possible the instance state could be inferred from the resource metadata | 09:48 |
| eglynn | ildikov_: ... I'd need to check whether we preserve the vm|task_state metadata, can't recall offhand | 09:48 |
| lsmola_ | ildikov_: just up or down | 09:49 |
| lsmola_ | ildikov_: so the samples are either 1 or 0 | 09:50 |
| lsmola_ | ildikov_: or missing | 09:50 |
| ildikov_ | eglynn: thanks, I will check that meter and the metadata too | 09:50 |
| nadya_ | eglynn: AFAIK, we store all metadata now | 09:50 |
| nadya_ | filtering may be broken a little bit, if structure of metadata is too complex | 09:51 |
| lsmola_ | nadya_: ok, cool | 09:51 |
| *** nadya__ has joined #openstack-ceilometer | 09:52 | |
| ildikov_ | lsmola_: for your charts, we need to handle the difference between 0 and missing and also we have to deal with how fine grained are the samples in a given time interval | 09:53 |
| nadya_ | nadya__, is it me again :)? | 09:53 |
| lsmola_ | ildikov_: not sure if that is needed | 09:54 |
| lsmola_ | ildikov_: if we are able to do sum of distinct samples | 09:54 |
| lsmola_ | ildikov_: 0 should be the same as missing sample | 09:55 |
| lsmola_ | ildikov_: if we are interested just in UP and DOWN | 09:55 |
| *** krast has quit IRC | 09:55 | |
| lsmola_ | ildikov_: yeah the period should be at least double of the sampling period, so there is always at least one sample | 09:56 |
| ildikov_ | lsmola_: you meant that missing also means down? | 09:56 |
| lsmola_ | ildikov_: maybe more, given there can be some latency | 09:56 |
| lsmola_ | ildikov_: well i am not sure about that | 09:56 |
| lsmola_ | ildikov_: maybe not | 09:57 |
| *** nadya__ has quit IRC | 09:57 | |
| lsmola_ | ildikov_: missing probably means that the instance has been removed, so it is not polled anymore | 09:57 |
| lsmola_ | ildikov_: which takes us to total_instance meter :-) | 09:57 |
| lsmola_ | ildikov_: cause that changes over time | 09:58 |
| ildikov_ | lsmola_: ok, then from your point of view it does not matter that the instance is just simply not running or removed? | 09:58 |
| lsmola_ | ildikov_: probably easier would be to show those two lines in chart | 09:58 |
| lsmola_ | ildikov_: not sure how to show percentage of VMs up | 09:58 |
| lsmola_ | ildikov_: in a line chart | 09:59 |
| eglynn | well the basic problems are (as I see it) ... | 09:59 |
| eglynn | 1. missing meter for a period doesn't necessarily imply instance was down | 09:59 |
| eglynn | 2. duplicate meter for a period doesn't necessarily imply more than one instance running | 09:59 |
| lsmola_ | ildikov_: well this was only to show, how many of them are up | 10:00 |
| lsmola_ | ildikov_: showing it in contrast of of the total vms is a different story, as i describe ^ | 10:00 |
| lsmola_ | eglynn: so 1. should be solved with big enough period right? | 10:01 |
| eglynn | lsmola_: yeap, with a "right-sized" period | 10:01 |
| ildikov_ | lsmola_: yep, sure, I understood that this is a different story | 10:02 |
| lsmola_ | ildikov_: though that would be nice to have too | 10:02 |
| ildikov_ | eglynn, lsmola_: can we make sure to always have the "right-sized" period? | 10:03 |
| lsmola_ | ildikov_: not really sure how to get timeseries of percentage of running machines | 10:03 |
| eglynn | ildikov_: we can't really as things stand, as the polling interval is embeded in the pipeline config and not directly visible to callers | 10:03 |
| ildikov_ | lsmola_:me neither | 10:03 |
| lsmola_ | ildikov_: it should depend n sampling period | 10:03 |
| eglynn | ildikov_: ... instead callers must infer that if they need to know | 10:03 |
| lsmola_ | eglynn: yeah, we will need to expose it | 10:04 |
| eglynn | lsmola_: yeap, that was the motivation for https://blueprints.launchpad.net/ceilometer/+spec/record-sample-periodicity | 10:04 |
| lsmola_ | eglynn: ok, cool | 10:04 |
| ildikov_ | eglynn: it looks really good, as we had some concerns about how to figure out, how many samples should we have for a given time interval | 10:06 |
| eglynn | cool | 10:06 |
| *** AMike has quit IRC | 10:11 | |
| ildikov_ | lsmola_: so as a first step, your chart would show the number of running instances for a project in given points of time in an interval, regarding to the interval configuration of this instance meter | 10:12 |
| *** sayali has joined #openstack-ceilometer | 10:15 | |
| openstackgerrit | liusheng proposed a change to openstack/ceilometer: check domain state before inspecting nics/disks https://review.openstack.org/68882 | 10:15 |
| openstackgerrit | liusheng proposed a change to openstack/ceilometer: check domain state before inspecting nics/disks https://review.openstack.org/68882 | 10:17 |
| ildikov_ | lsmola_: as for the paste, you've sent, we had the issue there with the average in the 1. point, regarding to the 0 or missing problem without the knowledge about the periodicity of that sample data | 10:17 |
| ildikov_ | lsmola_: I need to run now, I will get back to you, if we could figure out something for this | 10:19 |
| openstackgerrit | A change was merged to openstack/ceilometer: Move enable_acl and debug config to ceilometer.conf https://review.openstack.org/68166 | 10:22 |
| lsmola_ | ildikov_: ok, sounds great, thanks | 10:22 |
| openstackgerrit | A change was merged to openstack/ceilometer: Fix the Alarm documentation of Web API V2 https://review.openstack.org/68071 | 10:28 |
| *** xianghui has quit IRC | 10:28 | |
| *** EmilienM has quit IRC | 10:32 | |
| *** EmilienM has joined #openstack-ceilometer | 10:33 | |
| *** s2r2_ has quit IRC | 10:55 | |
| openstackgerrit | Ilya Tyaptin proposed a change to openstack/ceilometer: Add run pollsters on demand function to api https://review.openstack.org/66551 | 10:57 |
| openstackgerrit | A change was merged to openstack/ceilometer: StringIO compatibility for python3 https://review.openstack.org/67866 | 11:38 |
| *** _ruhe is now known as ruhe | 11:49 | |
| *** ildikov_ has quit IRC | 12:13 | |
| *** vrovachev has joined #openstack-ceilometer | 12:15 | |
| *** ildikov_ has joined #openstack-ceilometer | 12:25 | |
| openstackgerrit | A change was merged to openstack/python-ceilometerclient: Enable hacking H233 rule https://review.openstack.org/68531 | 12:28 |
| *** sayali_ has joined #openstack-ceilometer | 12:35 | |
| *** sayali has quit IRC | 12:38 | |
| openstackgerrit | Nadya Privalova proposed a change to openstack/ceilometer: Change meter_id to meter_name in docs https://review.openstack.org/68902 | 12:57 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 13:01 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 13:02 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 13:03 | |
| *** eglynn is now known as eglynn-lunch | 13:07 | |
| *** ruhe is now known as _ruhe | 13:07 | |
| *** prad has joined #openstack-ceilometer | 13:10 | |
| *** xianghui has joined #openstack-ceilometer | 13:10 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 13:11 | |
| *** _ruhe is now known as ruhe | 13:18 | |
| openstackgerrit | Nadya Privalova proposed a change to openstack/ceilometer: Change meter_id to meter_name in generated docs https://review.openstack.org/68902 | 13:23 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 13:24 | |
| *** jdob has joined #openstack-ceilometer | 13:27 | |
| *** vkodam has quit IRC | 13:28 | |
| *** thomasem has joined #openstack-ceilometer | 13:33 | |
| *** thomasem_ has joined #openstack-ceilometer | 13:41 | |
| *** openstackgerrit has quit IRC | 13:42 | |
| *** dhellmann has quit IRC | 13:42 | |
| vrovachev | Hi all, http://docs.openstack.org/developer/ceilometer/measurements.html#network-neutron. Where table for nova? | 13:43 |
| *** thomasem has quit IRC | 13:44 | |
| vrovachev | Guys who rules snippet in docs for nova, please return table. | 13:45 |
| *** dhellmann has joined #openstack-ceilometer | 13:49 | |
| eglynn-lunch | vrovachev: http://docs.openstack.org/developer/ceilometer/measurements.html#compute-nova | 13:49 |
| ildikov_ | eglynn-lunch, vrovachev: this link did not work for me either 5 minutes ago, only nova meters were disappeared from the Measurements page | 13:53 |
| ildikov_ | eglynn-lunch, vrovachev: currently my whole network connection seems to be dead, so I cannot check, what happened and it was permanent or just temorary | 13:54 |
| eglynn-lunch | ildikov_: odd, seems to work fine now for me ... | 13:55 |
| *** eglynn-lunch is now known as eglynn-really-lu | 13:55 | |
| vrovachev | page work fine | 13:56 |
| vrovachev | but table for nova | 13:56 |
| vrovachev | it is not | 13:56 |
| *** eglynn-really-lu is now known as eglynn | 13:57 | |
| eglynn | vrovachev: true that ... now that I've refreshed | 13:57 |
| ildikov_ | eglynn: Compute(Nova) subsection is still empty for me if I copy your link to my browser and press enter and wait 3 minutes to get the page downloaded :) | 13:58 |
| ildikov_ | eglynn: ok, fine, or not fine, but it's good that you see it too | 13:58 |
| * eglynn wonders if he broke the docs with this commit https://github.com/openstack/ceilometer/commit/b2fbf873 | 13:58 | |
| nadya_ | we need Sherlock | 13:59 |
| eglynn | (I was seeing cached content earlier maybe, dunno ... or not looking properly, more likely) | 13:59 |
| ildikov_ | eglynn: my git pull is suffering from the poor network access, so I cannot check now if it is that commit or no, as there was another patch for that part of the doc, I just do not know, if it was merged already or not | 14:00 |
| eglynn | ... /me feels a visit to the wayback-machine coming on ;) | 14:01 |
| nadya_ | btw, are you able to build the docs? | 14:01 |
| eglynn | meh! most recent cached copy is from June last year :( | 14:01 |
| eglynn | http://web.archive.org/web/20130626052717/http://docs.openstack.org/developer/ceilometer/measurements.html | 14:01 |
| ildikov_ | eglynn: I'm wondering, what is checked by the doc gate, if a change, which brokes the doc could pass | 14:03 |
| nadya_ | https://bugs.launchpad.net/openstack-ci/+bug/1268732 | 14:03 |
| ildikov_ | eglynn: it seems that it hasn't changed much besides your rate additions :) | 14:06 |
| eglynn | ildikov_: ... yeah here's the gate output for that patch | 14:06 |
| eglynn | http://docs-draft.openstack.org/57/66657/3/gate/gate-ceilometer-docs/d5017ab/doc/build/html/measurements.html#compute-nova | 14:06 |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 14:07 | |
| ildikov_ | eglynn: I've found the bug: /opt/stack/ceilometer/doc/source/measurements.rst:78: ERROR: Malformed table. | 14:07 |
| eglynn | ... and previous one to merge http://docs-draft.openstack.org/08/67608/3/gate/gate-ceilometer-docs/be503fa/doc/build/html/measurements.html#compute-nova | 14:07 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 14:09 | |
| eglynn | ildikov_: malformed in what sense I wonder, lines too wide now? | 14:09 |
| ildikov_ | eglynn: the 's' at the end of the request/s is too long | 14:10 |
| eglynn | ildikov_: ha-ha, well spotted ... fixing now! | 14:10 |
| ildikov_ | eglynn: you have to insert one more '=' in the header definition of that table I think | 14:10 |
| ildikov_ | eglynn: so I had my good deed for today ;) | 14:13 |
| eglynn | ildikov_: ... /me thanks your sharp eyes! ;) | 14:13 |
| eglynn | meh ... no more ircbot commenting on new gerrit proposals? | 14:14 |
| eglynn | https://review.openstack.org/68918 | 14:14 |
| * eglynn channels the ircbot ;) | 14:14 | |
| eglynn | ... bbiab | 14:15 |
| * ildikov_ looking :) | 14:15 | |
| gibi | it looks like we have a fun friday :) | 14:16 |
| ildikov_ | eglynn: have you run sphinx-build too? as I liked the previous version of this doc, after you corrected my first comments in it ;) | 14:17 |
| eglynn | ildikov_: nah, I just pushed it without building, let jenkins do that ... ;) | 14:18 |
| eglynn | ildikov_: ... not sure what you mean by "my first comments"? | 14:18 |
| ildikov_ | eglynn: to change Avg to Average | 14:20 |
| ildikov_ | eglynn: I will build it for you, I'm a sphinx-build expert, more or less :) | 14:21 |
| eglynn | ildikov_: ... a-ha, this change you mean? https://review.openstack.org/#/c/66658/3..4/doc/source/measurements.rst | 14:22 |
| *** xianghui has quit IRC | 14:22 | |
| eglynn | ildikov_: ... thanks! | 14:22 |
| eglynn | ... I'd keep that sphinx expertise under your hat if I was you! ;) | 14:23 |
| ildikov_ | eglynn: yes, when I criticized your grammar ;) | 14:24 |
| ildikov_ | eglynn: then I will not say that sphinx-build worked for me this time :) | 14:24 |
| eglynn | ildikov_: cool :) | 14:24 |
| vrovachev | people, help. I want to create alarm for metric "cpu_util" for instance with id "36f93dfa-1ba2-43a6-9392-9ea4b117babf". | 14:34 |
| vrovachev | My body: | 14:34 |
| vrovachev | "name":"alarmexample", | 14:34 |
| vrovachev | "alarm_actions":["http://site:8000/alarm"], | 14:34 |
| vrovachev | "description":"viva la ceilometer", | 14:34 |
| *** vrovachev has quit IRC | 14:34 | |
| *** vrovachev has joined #openstack-ceilometer | 14:35 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 14:41 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 14:41 | |
| eglynn | vrovachev: to capture the resource_id constraint you'll need ... | 14:43 |
| eglynn | "threshold_rule": {"query": [{"field": "resource_id", "type": "", "value": "36f93dfa-1ba2-43a6-9392-9ea4b117babf", "op": "eq"}], ...} | 14:43 |
| eglynn | but there's a bunch of other stuff needed too | 14:43 |
| eglynn | like the threshold value, the comparison operator, the statistic, the meter_name | 14:44 |
| eglynn | why not just use the CLI? | 14:44 |
| vrovachev | oh, in queue i use type: 'string'.... | 14:44 |
| vrovachev | i use tempest python client | 14:44 |
| vrovachev | https://review.openstack.org/#/c/55276/ | 14:45 |
| eglynn | vrovachev: a-ha, OK | 14:45 |
| eglynn | vrovachev: here's an example of what you need http://paste.openstack.org/show/61823/ | 14:48 |
| eglynn | (adjust the threshold, period etc. values for your own purposes ...) | 14:48 |
| vrovachev | eglynn: thank you very much! | 14:49 |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 14:51 | |
| *** ildikov_ has quit IRC | 15:10 | |
| *** tongli has joined #openstack-ceilometer | 15:18 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 15:19 | |
| tongli | @jd__, ping, | 15:30 |
| jd__ | tongli: pong | 15:30 |
| tongli | @jd__, ready to submit the first patch for the BP, and before I do that, can you give me an example of the use case? | 15:30 |
| tongli | @jd__, so that I can verify the implementation indeed does what BP wants. | 15:31 |
| jd__ | tongli: people wants to be alerted when Heat autoscale up or down a stack and therefore send a notification about that | 15:31 |
| tongli | @jd__, so the notification is quite specific, correct? | 15:32 |
| tongli | probably will be in a particular type with a particular payload, I would say? | 15:32 |
| jd__ | tongli: yes | 15:33 |
| tongli | @jd__, the question at hand is that how flexible do we want to make this in terms match up payload. | 15:33 |
| jd__ | tongli: a first version of this patch can omit I'd say | 15:33 |
| jd__ | otherwise we can mimick the matching mechanism used on the threshold alarm | 15:34 |
| tongli | the operator on the type, I was thinking having eq, ne, startswith, endswith. | 15:34 |
| tongli | i c. | 15:34 |
| tongli | r u ok with the operators for matching the type? | 15:34 |
| EmilienM | jd__: hey, would you mind if I add you as reviewer of ceilometer-agent-notification support in puppet-ceilometer? | 15:38 |
| tongli | @jd__, r u ok with the operators (eq, ne, startswith, endswith) for matching notification type? | 15:38 |
| jd__ | EmilienM: go ahead | 15:39 |
| EmilienM | jd__: thanks! > https://review.openstack.org/#/c/68940/ | 15:39 |
| jd__ | tongli: what about having a simple wildcard support like we do in pipeline? foo.* | 15:39 |
| tongli | @jd__, i c, use regular expression. | 15:40 |
| tongli | in that case, we do not have to use operators. | 15:40 |
| tongli | simply say match, ok. | 15:40 |
| tongli | I think that is a good idea. | 15:40 |
| jd__ | not regex, just wildcard really | 15:40 |
| jd__ | fnmatch is used in ceilometer.pipeline, I think it's simple and good enough | 15:40 |
| tongli | @jd__, that works. | 15:40 |
| jd__ | tongli: you may still want to have "matches" and "do not matches" (eq/ne) | 15:41 |
| EmilienM | jd__: sorry, i fix a header stuff in the last patchset. can you +1 again? You were so fast | 15:41 |
| jd__ | just a way to negate | 15:41 |
| jd__ | tongli: we use ! as an operator for that in pipeline, maybe you can reuse that too? even share the code | 15:41 |
| tongli | @jd__, yeah, eq, ne operator. | 15:41 |
| tongli | yeah, I will take a look at that. | 15:41 |
| tongli | @jd__, good ideas, thanks. | 15:42 |
| tongli | the patch will be there soon. thanks @jd__. | 15:42 |
| *** gordc has joined #openstack-ceilometer | 16:07 | |
| *** openstackgerrit has joined #openstack-ceilometer | 16:12 | |
| openstackgerrit | John Herndon proposed a change to openstack/ceilometer: Update dev docs to include notification-agent https://review.openstack.org/67252 | 16:14 |
| *** vrovachev has left #openstack-ceilometer | 16:18 | |
| prad | jaypipes: jd__ : would like to know how we want to proceed on https://review.openstack.org/#/c/66746/ .. so if those are not really metering anything, wonder why do they even show up as part of the metering data | 16:18 |
| *** sayali_ has quit IRC | 16:36 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 16:37 | |
| *** sayali_ has joined #openstack-ceilometer | 16:37 | |
| *** HT_Sergio has joined #openstack-ceilometer | 16:53 | |
| *** ruhe is now known as _ruhe | 17:04 | |
| *** xmltok has joined #openstack-ceilometer | 17:04 | |
| * eglynn would appreciate a sanity check on a kinda wacky idea ... | 17:12 | |
| eglynn | basically, *parameterized* statistical aggregate functions | 17:13 |
| eglynn | so currently we support (sum, avg, max, min, count) | 17:13 |
| eglynn | none of which are externally parameterized | 17:13 |
| jaypipes | prad: I had the same question :) | 17:13 |
| eglynn | ... i.e. the results are completely determined by the sample data | 17:13 |
| eglynn | but here are two cases that might benefit from parameterization ... | 17:14 |
| jaypipes | eglynn: not quite following you... | 17:14 |
| jaypipes | by "externally parameterized".. what do you mean? | 17:14 |
| eglynn | jaypipes: an example might make it clearer: | 17:15 |
| jaypipes | k | 17:15 |
| *** thomasem_ has left #openstack-ceilometer | 17:15 | |
| eglynn | example 1. free-form quantiles | 17:15 |
| eglynn | (i.e. the capability to request median, OR p99, OR all 10 deciles, etc.) | 17:15 |
| eglynn | ... here the quantile calculation would be parameterized by the cut-off point(s) | 17:15 |
| eglynn | say ... 0.5 for median | 17:16 |
| eglynn | or [0.1,0.2,0.3,...0.9,1.0] for deciles | 17:16 |
| eglynn | i.e. the caller into the statistics API would provide those parameter(s) | 17:16 |
| jaypipes | ok, that makes more sense. | 17:16 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 17:17 | |
| eglynn | another example would be distinct counting based on some attribute | 17:17 |
| eglynn | e.g. the ability to ask for the number of distinct resource_ids in each grouped-by bucket | 17:17 |
| eglynn | (so here the parameter would be "resource_id" to identify the attr for which we want a distinct count on) | 17:18 |
| jaypipes | eglynn: this functionality veers a little too close into existing OLAP solutions... :) | 17:18 |
| eglynn | jaypipes: yeap, I hear ya | 17:18 |
| jaypipes | devil's advocacy. | 17:18 |
| eglynn | the disctinct counting would be one way of addressing the horizon requirement for direct access to per-tenant resource counts | 17:19 |
| eglynn | i.e. GET /v2/meter/instance/statistics?groupby=project_id&distinct=resource_id&period=600 | 17:20 |
| jaypipes | eglynn: or alternately GET /v2/projects/meters?group_by=instance_id&start_timestamp=now-600s | 17:22 |
| jaypipes | sorry... | 17:22 |
| jaypipes | eglynn: or alternately GET /v2/projects/$project_id/meters?group_by=instance_id&start_timestamp=now-600s | 17:22 |
| eglynn | jaypipes: yeah, we kinda moved away from slicing on project id with the transition from the v1 to the v2 ceilo API | 17:24 |
| eglynn | plus the allinone queryability is I think important for the dashboard folks | 17:25 |
| jaypipes | eglynn: yes, I see that. no worries. | 17:25 |
| eglynn | yeah I need to think about some more | 17:25 |
| jaypipes | eglynn: unfortunately, I was a bit lost in ops-land for 18-20 months :( Just now getting back into the swing of upstream contribution after a long walk in the desert. | 17:25 |
| jaypipes | so please forgive my learning curve! | 17:26 |
| eglynn | jaypipes: absolutely, I'm sure your adventures in ops-land will pay off big-style now ... nothing like real experience of running at scale | 17:27 |
| jaypipes | eglynn: yes, that's actually part of the reason I want to contribute to ceilometer and neutron. those are the two projects that we found most difficult to scale and I'm hoping to assist in improving both projects in that manner. | 17:28 |
| eglynn | jaypipes: great, your ideas on scaling ceilo will be much appreciated | 17:28 |
| jaypipes | eglynn: ty. will try my best. will try not to piss jd__ off too much with my newbie review questions ;) | 17:29 |
| eglynn | ... don't worry about jd__, he's got a thick skin ;) | 17:29 |
| jaypipes | heh | 17:30 |
| jd__ | :) | 17:31 |
| jd__ | eglynn: I think that's for you https://twitter.com/noggin143/status/426770246834196480 :) | 17:38 |
| openstackgerrit | John Herndon proposed a change to openstack/python-ceilometerclient: Update client to display data type of traits https://review.openstack.org/67224 | 17:39 |
| *** SergeyLukjanov is now known as SergeyLukjanov_a | 17:42 | |
| *** SergeyLukjanov_a is now known as SergeyLukjanov_ | 17:43 | |
| *** yassine has quit IRC | 17:46 | |
| eglynn | jd__: a-ha, cool ... /me must get on twitter and catch up with the 21st century | 17:51 |
| jd__ | :-) | 17:51 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 18:12 | |
| *** eglynn is now known as eglynn-afk | 18:20 | |
| *** dhellmann is now known as dhellmann_ | 18:23 | |
| *** ildikov has joined #openstack-ceilometer | 18:29 | |
| *** nadya__ has joined #openstack-ceilometer | 18:40 | |
| *** ildikov_ has joined #openstack-ceilometer | 18:42 | |
| openstackgerrit | Nadya Privalova proposed a change to openstack/ceilometer: Fix work with metadata in HBase https://review.openstack.org/68641 | 18:55 |
| *** julienvey_ has joined #openstack-ceilometer | 19:04 | |
| *** flwang has quit IRC | 19:14 | |
| *** xmltok has quit IRC | 19:17 | |
| *** flwang has joined #openstack-ceilometer | 19:20 | |
| *** dhellmann_ is now known as dhellmann | 19:24 | |
| *** sayali_ has quit IRC | 19:51 | |
| *** nadya__ has quit IRC | 19:56 | |
| *** xmltok has joined #openstack-ceilometer | 20:04 | |
| *** nadya__ has joined #openstack-ceilometer | 20:11 | |
| *** prad has quit IRC | 20:12 | |
| *** HT_Sergio has quit IRC | 20:13 | |
| *** rwsu has quit IRC | 20:14 | |
| *** rwsu has joined #openstack-ceilometer | 20:19 | |
| *** ok_delta has joined #openstack-ceilometer | 20:23 | |
| *** ildikov has quit IRC | 20:25 | |
| *** pym has joined #openstack-ceilometer | 20:27 | |
| pym | views? | 20:29 |
| openstackgerrit | Nadya Privalova proposed a change to openstack/ceilometer: Fix work with metadata in HBase https://review.openstack.org/68641 | 20:36 |
| *** _ruhe is now known as ruhe | 20:53 | |
| openstackgerrit | Nadya Privalova proposed a change to openstack/ceilometer: Fix work with metadata in HBase https://review.openstack.org/68641 | 21:08 |
| *** nadya__ has quit IRC | 21:09 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Fix wrong doc string for meter type https://review.openstack.org/68582 | 21:21 |
| *** kwhitney has quit IRC | 21:33 | |
| *** ruhe is now known as _ruhe | 21:33 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 21:39 | |
| *** jdob has quit IRC | 21:43 | |
| *** gordc has quit IRC | 21:56 | |
| openstackgerrit | A change was merged to openstack/python-ceilometerclient: Avoid discarding alarm-threshold-create --query option https://review.openstack.org/68637 | 22:00 |
| *** julienvey_ has quit IRC | 22:01 | |
| *** julienvey_ has joined #openstack-ceilometer | 22:04 | |
| *** xmltok has quit IRC | 22:10 | |
| *** xmltok has joined #openstack-ceilometer | 22:10 | |
| *** tongli has quit IRC | 22:26 | |
| *** eglynn-afk has quit IRC | 22:28 | |
| openstackgerrit | Jenkins proposed a change to openstack/ceilometer: Updated from global requirements https://review.openstack.org/68237 | 22:33 |
| openstackgerrit | Jenkins proposed a change to openstack/python-ceilometerclient: Updated from global requirements https://review.openstack.org/69039 | 22:40 |
| *** _cjones_ has joined #openstack-ceilometer | 22:42 | |
| _cjones_ | Any Ceilometer people here today for a general question? I'm curious as to the status of Alarm/Alerting functionality in Ceilometer. | 23:12 |
| *** julienvey_ has quit IRC | 23:33 | |
| openstackgerrit | ZhiQiang Fan proposed a change to openstack/ceilometer: Use explicit http error code for api v2 https://review.openstack.org/68775 | 23:47 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!