| *** zul has quit IRC | 00:17 | |
| *** fnaval_ has quit IRC | 00:18 | |
| *** zul has joined #openstack-metering | 00:26 | |
| *** flwang has quit IRC | 00:43 | |
| *** herndon has joined #openstack-metering | 00:47 | |
| *** fnaval_ has joined #openstack-metering | 00:48 | |
| *** llu has quit IRC | 01:04 | |
| *** nosnos has joined #openstack-metering | 01:22 | |
| *** flwang has joined #openstack-metering | 01:49 | |
| *** fandikurnia01 has joined #openstack-metering | 01:53 | |
| fandikurnia01 | hello all | 01:54 |
|---|---|---|
| fandikurnia01 | i am install openstack using mirantis fuel | 01:54 |
| fandikurnia01 | how to enable ceilometer service ? | 01:55 |
| fandikurnia01 | on controller | 01:55 |
| fandikurnia01 | thanks | 01:55 |
| *** thomasm has joined #openstack-metering | 01:59 | |
| *** d34dh0r53 has joined #openstack-metering | 01:59 | |
| *** d34dh0r53 has quit IRC | 02:01 | |
| *** herndon has quit IRC | 02:01 | |
| *** herndon has joined #openstack-metering | 02:02 | |
| *** llu has joined #openstack-metering | 02:19 | |
| openstackgerrit | litong01 proposed a change to openstack/ceilometer: refactor db2 get_meter_statistics method to support mongodb and db2 https://review.openstack.org/46175 | 02:21 |
| *** nati_ueno has quit IRC | 02:22 | |
| *** haomeng has joined #openstack-metering | 02:43 | |
| *** nosnos has quit IRC | 02:57 | |
| *** anteaya__ is now known as anteaya | 02:57 | |
| *** nosnos has joined #openstack-metering | 02:58 | |
| *** fnaval__ has joined #openstack-metering | 03:01 | |
| openstackgerrit | gordon chung proposed a change to openstack/ceilometer: Pecan assuming meter names are extensions https://review.openstack.org/46143 | 03:01 |
| *** fnaval_ has quit IRC | 03:03 | |
| openstackgerrit | Terri Yu proposed a change to openstack/ceilometer: Add group by statistics examples in API v2 docs https://review.openstack.org/46181 | 03:06 |
| *** fnaval__ has quit IRC | 03:27 | |
| *** fnaval_ has joined #openstack-metering | 03:27 | |
| *** anteaya has quit IRC | 03:33 | |
| *** shang has quit IRC | 03:37 | |
| openstackgerrit | A change was merged to openstack/ceilometer: validate counter_type when posting samples https://review.openstack.org/45344 | 03:53 |
| *** SergeyLukjanov has joined #openstack-metering | 04:13 | |
| *** SergeyLukjanov has quit IRC | 04:14 | |
| *** SergeyLukjanov has joined #openstack-metering | 04:20 | |
| *** SergeyLukjanov has quit IRC | 04:30 | |
| *** SergeyLukjanov has joined #openstack-metering | 04:33 | |
| *** herndon has quit IRC | 04:38 | |
| *** SergeyLukjanov has quit IRC | 04:48 | |
| *** SergeyLukjanov has joined #openstack-metering | 04:57 | |
| *** SergeyLukjanov has quit IRC | 04:57 | |
| *** fnaval_ has quit IRC | 05:04 | |
| *** boris-42 has joined #openstack-metering | 05:22 | |
| openstackgerrit | Lianhao Lu proposed a change to openstack/ceilometer: Use global openstack requirements https://review.openstack.org/46189 | 05:30 |
| *** mmcardle has joined #openstack-metering | 05:41 | |
| *** SergeyLukjanov has joined #openstack-metering | 05:42 | |
| openstackgerrit | Haomeng,Wang proposed a change to openstack/ceilometer: Change resource.resource_metadata column to Text https://review.openstack.org/45064 | 05:50 |
| openstackgerrit | Jenkins proposed a change to openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/46194 | 06:02 |
| *** nosnos has quit IRC | 06:17 | |
| *** mmcardle has quit IRC | 06:17 | |
| *** nosnos has joined #openstack-metering | 06:24 | |
| *** SergeyLukjanov has quit IRC | 06:43 | |
| openstackgerrit | Lianhao Lu proposed a change to openstack/ceilometer: Correrctly output the sample content in file publisher https://review.openstack.org/46202 | 06:50 |
| lsmola | jd__, hello | 06:56 |
| lsmola | jd__, could you please look whether these are relevant bugs, or I am just missing something? | 06:57 |
| lsmola | jd__, https://bugs.launchpad.net/ceilometer/+bug/1223829 , https://bugs.launchpad.net/ceilometer/+bug/1224298 | 06:57 |
| lsmola | jd__, thank you very much | 06:58 |
| sileht | lsmola, bug/1224298 is relevant | 07:01 |
| sileht | I think bug/1223829 is not | 07:01 |
| sileht | the project_id/owner_id at the root of the json, are the alarm owner, this isn't modifiable | 07:01 |
| sileht | but If you want to set a alarm on the a particular project_id, you can: | 07:02 |
| sileht | use matching_metadata : { 'project_id': '<UUID>' } | 07:02 |
| sileht | lsmola, make sense ? | 07:02 |
| sileht | lsmola, but we are currenlty change a bit the alarm definition to make it more understandable | 07:03 |
| lsmola | sileht, ok | 07:03 |
| lsmola | sileht, so we should probably remove project_id and user_id from https://github.com/openstack/python-ceilometerclient/blob/master/ceilometerclient/v2/alarms.py#L41 | 07:05 |
| lsmola | sileht, cool, are those matching_metadata documented somewhere? | 07:05 |
| sileht | lsmola, no because when a admin retrieve a alarm this is interesting to know the owner | 07:05 |
| openstackgerrit | Harri Hämäläinen proposed a change to openstack/ceilometer: Catch exceptions from nova client in poll_and_publish. https://review.openstack.org/46203 | 07:05 |
| lsmola | sileht, cause i didn't notice there is some functionality connected to metadata | 07:06 |
| sileht | lsmola, I'm currently changing the matching_meta object by the Query object (already documented) | 07:06 |
| sileht | matching_meta is a dict of key/value to match any sample field | 07:06 |
| *** eglynn has quit IRC | 07:07 | |
| lsmola | sileht, oh, interesting, I totally missed that | 07:07 |
| lsmola | sileht, thought for a while I can set alarms only for the whole project, cool, will try this out | 07:08 |
| lsmola | sileht, actually this is filter only from creation, right ? https://github.com/openstack/python-ceilometerclient/blob/master/ceilometerclient/v2/alarms.py#L41 | 07:08 |
| lsmola | sileht, from/for | 07:08 |
| sileht | lsmola, no all sample fields | 07:09 |
| lsmola | sileht, it is not filtering what attributes the fetched object will have | 07:09 |
| lsmola | sileht, so I would remove that from the array, cause otherwise it is confusing, seem like it accepts user_id and project_id, but it is not | 07:12 |
| sileht | lsmola, on ceilometerclient you are right | 07:13 |
| lsmola | sileht, yes, I think I mention it in the bug also | 07:14 |
| lsmola | sileht, i will change it to that option then | 07:14 |
| *** shardy_afk is now known as shardy | 07:15 | |
| lsmola | sileht, hmm trying to find some description of matching_metadata, is there somewhere some more verbose docs? Could be maybe good to add more verbose description to V2 docs | 07:15 |
| sileht | lsmola, it will change to be Query object very soon | 07:15 |
| lsmola | sileht, maybe with some real example | 07:15 |
| lsmola | sileht, ok | 07:16 |
| lsmola | sileht, is there a bug or blueprint for this? | 07:16 |
| lsmola | sileht, I am obsessed with dependency tracking :-D | 07:16 |
| sileht | lsmola, I work on https://blueprints.launchpad.net/ceilometer/+spec/alarming-logical-combination | 07:17 |
| sileht | lsmola, the final v2 API alarm for havana https://wiki.openstack.org/wiki/Ceilometer/blueprints/alarm-api | 07:17 |
| sileht | lsmola, I just see, this is for horizon cool :) | 07:18 |
| openstackgerrit | Terri Yu proposed a change to openstack/ceilometer: Add timestamp filtering cases in storage tests https://review.openstack.org/46206 | 07:19 |
| lsmola | sileht, oh yeah, I am already tracking those, cool | 07:20 |
| lsmola | sileht, I have changed the description of https://bugs.launchpad.net/ceilometer/+bug/1223829 , now it should be valid bug, right? | 07:22 |
| openstackgerrit | A change was merged to openstack/ceilometer: Make the Swift-related doc more explicit https://review.openstack.org/45834 | 07:24 |
| sileht | lsmola, perfect | 07:38 |
| openstackgerrit | Harri Hämäläinen proposed a change to openstack/ceilometer: Catch exceptions from nova client in poll_and_publish https://review.openstack.org/46203 | 07:38 |
| lsmola | sileht, cool, thank you very much | 07:39 |
| *** boris-42 has quit IRC | 07:55 | |
| *** Ruetobas has quit IRC | 08:04 | |
| *** SergeyLukjanov has joined #openstack-metering | 08:09 | |
| *** lexx has joined #openstack-metering | 08:26 | |
| *** mmcardle has joined #openstack-metering | 08:30 | |
| *** eglynn has joined #openstack-metering | 08:32 | |
| *** neumerance has joined #openstack-metering | 08:41 | |
| neumerance | I need some help | 08:41 |
| neumerance | Yoh Dhellman. | 08:42 |
| *** taplax has joined #openstack-metering | 09:01 | |
| *** mmcardle has left #openstack-metering | 09:08 | |
| *** vvechkanov2 has quit IRC | 09:12 | |
| *** Joy has joined #openstack-metering | 09:21 | |
| *** boris-42 has joined #openstack-metering | 09:22 | |
| *** Joy has quit IRC | 09:23 | |
| *** Joy has joined #openstack-metering | 09:25 | |
| *** flwang has quit IRC | 09:51 | |
| *** lexx has quit IRC | 09:54 | |
| openstackgerrit | Julien Danjou proposed a change to openstack/ceilometer: doc: fix storage backend features status https://review.openstack.org/44453 | 09:56 |
| openstackgerrit | A change was merged to openstack/ceilometer: Update config generator https://review.openstack.org/45885 | 10:15 |
| openstackgerrit | A change was merged to openstack/ceilometer: Include auth_token middleware in sample config https://review.openstack.org/45886 | 10:16 |
| openstackgerrit | A change was merged to openstack/ceilometer: add tests for _query_to_kwargs func https://review.openstack.org/43796 | 10:21 |
| openstackgerrit | A change was merged to openstack/ceilometer: Add docstrings to some methods https://review.openstack.org/46141 | 10:25 |
| openstackgerrit | A change was merged to openstack/ceilometer: Fix handling of bad paths in Swift middleware https://review.openstack.org/45983 | 10:28 |
| *** SergeyLukjanov has quit IRC | 10:36 | |
| *** SergeyLukjanov has joined #openstack-metering | 10:38 | |
| *** SergeyLukjanov has quit IRC | 10:54 | |
| *** flwang has joined #openstack-metering | 10:55 | |
| *** eglynn is now known as eglynn-lunch | 11:03 | |
| openstackgerrit | Harri Hämäläinen proposed a change to openstack/ceilometer: Catch exceptions from nova client in poll_and_publish https://review.openstack.org/46203 | 11:06 |
| *** SergeyLukjanov has joined #openstack-metering | 11:07 | |
| *** shang has joined #openstack-metering | 11:09 | |
| *** neumerance has quit IRC | 11:20 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Handle missing libvirt vnic targets! https://review.openstack.org/45444 | 11:20 |
| *** fandikurnia01 has quit IRC | 11:28 | |
| openstackgerrit | Cyril Roelandt proposed a change to openstack/python-ceilometerclient: Help messages: specify which options are required https://review.openstack.org/45870 | 11:32 |
| *** chuckieb has joined #openstack-metering | 11:44 | |
| *** shang has quit IRC | 11:55 | |
| *** annegentle has quit IRC | 12:08 | |
| *** changbl has quit IRC | 12:27 | |
| *** shang has joined #openstack-metering | 12:30 | |
| *** fandikurnia01 has joined #openstack-metering | 12:40 | |
| *** fandikurnia01 has quit IRC | 12:49 | |
| *** herndon has joined #openstack-metering | 12:53 | |
| *** annegentle has joined #openstack-metering | 12:57 | |
| *** anteaya has joined #openstack-metering | 12:58 | |
| *** bpokorny has joined #openstack-metering | 12:58 | |
| *** annegentle is now known as annegentle_away | 13:02 | |
| *** sandywalsh has quit IRC | 13:08 | |
| *** gordc has joined #openstack-metering | 13:12 | |
| *** shang_ has joined #openstack-metering | 13:14 | |
| *** zul has quit IRC | 13:18 | |
| *** zul has joined #openstack-metering | 13:18 | |
| *** sandywalsh has joined #openstack-metering | 13:20 | |
| *** Ruetobas has joined #openstack-metering | 13:21 | |
| *** Ruetobas has quit IRC | 13:25 | |
| *** nosnos has quit IRC | 13:26 | |
| *** lnxnut has joined #openstack-metering | 13:29 | |
| *** Ruetobas has joined #openstack-metering | 13:32 | |
| thomasm | Hey all! | 13:33 |
| *** eglynn-lunch is now known as eglynn | 13:35 | |
| *** Ruetobas has quit IRC | 13:36 | |
| *** changbl has joined #openstack-metering | 13:38 | |
| *** Ruetobas has joined #openstack-metering | 13:44 | |
| lnxnut | hey | 13:44 |
| *** shang_ has quit IRC | 13:48 | |
| lsmola | eglynn, hello, do you have a time for a quick question? | 13:49 |
| eglynn | lsmola: hi, sure! | 13:50 |
| eglynn | lsmola: shoot :) | 13:50 |
| *** lnxnut has quit IRC | 13:50 | |
| lsmola | eglynn, I can't find alarm-history in ceilometer client, I guess it is not yet implemented right? | 13:51 |
| eglynn | lsmola: exactomundo! | 13:51 |
| *** shang has quit IRC | 13:51 | |
| lsmola | eglynn, I have created a blueprint for that https://blueprints.launchpad.net/python-ceilometerclient/+spec/alarm-audit-client-api | 13:51 |
| eglynn | lsmola: (it's a pretty new feature, so the client hasn't caught up yet) | 13:51 |
| *** shang has joined #openstack-metering | 13:51 | |
| eglynn | lsmola: great, thanks! | 13:51 |
| lsmola | eglynn, would you have time for that? or should I try to implement it? | 13:52 |
| lsmola | eglynn, :-) | 13:52 |
| openstackgerrit | Dan Prince proposed a change to openstack/ceilometer: Handle inst not found exceptions in pollsters https://review.openstack.org/45802 | 13:52 |
| eglynn | lsmola: well it's on my list, but not at the top of the priority stack as yet ... so if you wanted to take a crack it, knock yourself out! | 13:52 |
| eglynn | lsmola: (otherwise I'll get to it eventually ... ;)) | 13:53 |
| lsmola | eglynn, ok cool, I am implementing the API on Horizon side, so once I get to it, I might do it :-) | 13:53 |
| lsmola | eglynn, another question is about a query | 13:53 |
| eglynn | lsmola: great, thanks! (lemme know if you do start on it ...) | 13:53 |
| eglynn | lsmola: sure | 13:53 |
| lsmola | eglynn, I do see that the type should serve for queries regarding state change, though I am not sure how this works | 13:54 |
| eglynn | lsmola: one sec | 13:55 |
| lsmola | eglynn, can I for instance list all alarms, that were at least once in alarm state in e.g. last week? | 13:55 |
| thomasm | eglynn, Hey, what's the criteria for a bug fix to be backported to Grizzly as well? I think that resource metadata bug that just got into RC1 may also be present in Grizzly. | 13:55 |
| thomasm | bug fix that just got into RC1* | 13:55 |
| thomasm | Ah, sorry. I'll ask again when y'all are done conversing. Don't want to confuse everyone. | 13:56 |
| eglynn | thomasm: do you mean what are appropriate fixes for backporting to stable/grizzly? https://wiki.openstack.org/wiki/StableBranch#Appropriate_Fixes | 13:56 |
| thomasm | yeah | 13:57 |
| thomasm | Oh, sweet. Thanks | 13:57 |
| thomasm | So already a small bit of the fix is out, schema change in HBase | 13:59 |
| openstackgerrit | gordon chung proposed a change to openstack/ceilometer: Pecan assuming meter names are extensions https://review.openstack.org/46143 | 14:03 |
| thomasm | eglynn, thanks - I'm not entirely sure it'd be a good idea with the pretty big get_resources query change and what-not. | 14:04 |
| eglynn | lsmola: sorry, back again ... | 14:05 |
| eglynn | lsmola: something like /v2/alarms/ALARM_ID/history?q.op=ge&q.op=le&q.op=eq&q.value=NOW&q.value=ONE_WEEK_AGO&q.value=state%20transition&q.field=timestamp&q.field=timestamp&q.field=type | 14:05 |
| eglynn | will give you all the state transitions within the last week | 14:05 |
| eglynn | then you'd have to filter on the details field for transitions into alarm (as opposed to transitions into OK or insufficient data) | 14:06 |
| eglynn | thomasm: normally the preference is for stable fixes to be self-contained and low-risk | 14:06 |
| eglynn | thomasm: ... but obviously that's the ideal, and the reality can be different | 14:07 |
| eglynn | thomasm: is the hbase change large and/or risky? | 14:07 |
| thomasm | eglynn, For sure. The bug is actually that the latest resource representation (from a sample) wasn't being returned from the resources API. | 14:07 |
| thomasm | eglynn, It's for all DB drivers | 14:07 |
| eglynn | thomasm: is the backport already proposed on gerrit, or something you're still actively working on? | 14:08 |
| thomasm | eglynn, I just tagged it as grizzly-backport-potential while I was making the fix for Havana-3. https://bugs.launchpad.net/ceilometer/+bug/1208547 | 14:09 |
| * eglynn looking ... | 14:10 | |
| *** herndon has quit IRC | 14:10 | |
| eglynn | thomasm: well at least it doesn't need a migration for sqlalchemy | 14:14 |
| eglynn | thomasm: (that would be a deal-breaker for stable, as we haven't pre-reserved any sa-migrate index slots for stable/grizzly) | 14:14 |
| eglynn | thomasm: however the backport may well be non-trivial as the storage driver code has had a fair bit of churn | 14:15 |
| thomasm | eglynn, That's what I was thinking | 14:15 |
| thomasm | eglynn, Like, it wasn't just a simple order by for SQLAlchemy, because there's undefined behavior when pulling non-aggregated values. | 14:15 |
| thomasm | MongoDB needed to be sorted on timestamp DESC, then it gave the latest | 14:16 |
| eglynn | thomasm: if you were interested in trying a cherry-pick onto stable, to see how messy it gets, that would be useful info | 14:16 |
| thomasm | eglynn, Yeah, I need to get my devstack set up with stable and give it a go. | 14:17 |
| eglynn | thomasm: cool | 14:17 |
| lsmola | eglynn, ok meaning the grouping and filtering have to be done in python, ok I will try it | 14:17 |
| thomasm | eglynn, I would only be able to address the drivers I can actually get running, though. Heh? DB2 is out unless litong wants to help me out there. | 14:17 |
| lsmola | eglynn, can I set you as the approver of the https://blueprints.launchpad.net/python-ceilometerclient/+spec/alarm-audit-client-api ? | 14:17 |
| eglynn | lsmola: done :) | 14:18 |
| lsmola | eglynn, cool, thank you very much | 14:19 |
| thomasm | HBase is out because of schema changes anyway. | 14:19 |
| gordc | thomasm: db2 isn't in grizzly so no need to worry about that. :) | 14:19 |
| thomasm | gordc, Oh my? I don't know why I thought it was. | 14:19 |
| eglynn | lsmola: do you mean group by alarm ID? ... well the basic API structure requires that the history is queried per-alarm anyway | 14:19 |
| thomasm | <--- needs more coffee | 14:20 |
| thomasm | gordc, Thanks :P | 14:20 |
| eglynn | lsmola: ... whereas the filtering on target state being alarm being needed on the client side, yep | 14:20 |
| gordc | thomasm: that said, i'm not sure you'd be able to backport mongo either... your patch uses relies on aggregate and i'm pretty sure we didn't/couldn't use aggregate in grizzly. | 14:21 |
| gordc | thomasm: but yeah, have a look if you have time. i need some coffee too so i could be talking non-sense right now. | 14:22 |
| lsmola | eglynn, I will get back all the alarm samples right? And I have to find the ones that had Alarm transitions | 14:23 |
| eglynn | lsmola: all the change events for a *particular* alarm that involve state transitions (as opposed to creation, deletion, or rule changes) | 14:24 |
| *** annegentle_away is now known as annegentle | 14:24 | |
| lsmola | eglynn, oh I see, it is per alarm | 14:24 |
| eglynn | lsmola: /v2/alarms/ALARM_ID/history requests the change events for the alarm with ALARM_ID only | 14:25 |
| *** fnaval_ has joined #openstack-metering | 14:25 | |
| eglynn | lsmola: yep, exactly | 14:25 |
| lsmola | eglynn, how hard would it be to make a query over multiple alarms? | 14:25 |
| eglynn | 14:25 | |
| lsmola | eglynn, it is something similar as sample-api for samples | 14:25 |
| eglynn | lsmola: samples are serverd out per-meter, history events are per-alarm, that's just the way the API is constructed currently | 14:27 |
| lsmola | eglynn, it will be very needed when i want to display e.g. health map of nova instances, wchich would obtain Alarms by several meter_names, group them by some metadata (nova instance id) and give me back number of Alarm transitions | 14:28 |
| lsmola | eglynn, yeah jd__ is preparing sample-api, that will be query over multiple samples | 14:28 |
| lsmola | eglynn, https://blueprints.launchpad.net/ceilometer/+spec/sample-api | 14:29 |
| eglynn | lsmola: well we would need to extend the alarm history API to be capable of querying over multiple alarms ... it's just not supported currently | 14:29 |
| lsmola | eglynn, I guess alarm-history could use something similar | 14:29 |
| eglynn | lsmola: the other thing to note is that alarm rules can change over time | 14:29 |
| lsmola | eglynn, yes, should i create a blueprint for this? | 14:29 |
| eglynn | lsmola: so for example I could have an alarm X that includes now includes matching metadata that identifies instance Y | 14:30 |
| lsmola | eglynn, hmm, we could group alarms by more attributes then | 14:30 |
| eglynn | lsmola: it could undergo a state transition on that basis, but then be changed to refer to instance Z | 14:31 |
| lsmola | eglynn, what information does the sample contain? | 14:31 |
| eglynn | lsmola: when you say sample, do you mean alarm history event? | 14:32 |
| lsmola | eglynn, if it has only reference to the alarm, that could be a problem | 14:32 |
| lsmola | eglynn, yes | 14:32 |
| *** yanghy has joined #openstack-metering | 14:33 | |
| lsmola | eglynn, if the event actually store the metadata + the conditions, I could theoretically group by that | 14:33 |
| eglynn | lsmola: it currently just includes ... event_id, alarm_id, type, detail, user_id, project_id, on_behalf_of | 14:34 |
| lsmola | eglynn, lets say it would be enough for me to know, how many alerts of some meters, did the instance have over some time | 14:34 |
| eglynn | lsmola: again that could be tricky, as the instance can be identified in multiple ways | 14:35 |
| eglynn | lsmola: e.g. an alarm could refer to it directly by resource_id | 14:35 |
| lsmola | eglynn, the instance should be probably identified in matching_metadata, right? | 14:35 |
| eglynn | lsmola: or indirectly via some other metadata | 14:35 |
| eglynn | lsmola: for example, heat autoscaling alarms identify groups of instances on the basis of their user metadata | 14:36 |
| lsmola | eglynn, hmm | 14:36 |
| eglynn | lsmola: so just knowing the instance ID is not enough to find all matching alarm | 14:36 |
| eglynn | s | 14:37 |
| lsmola | eglynn, though then the alarm is not tied directly to one resource right? | 14:37 |
| eglynn | lsmola: exactly | 14:37 |
| lsmola | eglynn, I think it could be usable | 14:37 |
| eglynn | lsmola: for the "wide dimensioning" case, a single alarm may refer to statistics averaged over many instances | 14:38 |
| eglynn | lsmola: e.g. all instances booted from a certain image, all instances with certain user metadata set etc. | 14:38 |
| lsmola | eglynn, if we want a health map chart showing resources, we must have alarms tied to resources | 14:38 |
| lsmola | eglynn, then we could show health map on resource aggregate level, etc. | 14:38 |
| eglynn | lsmola: alarms are currently not neccessarily tied to individual resources | 14:39 |
| lsmola | eglynn, yes, though I would filter only those that are, if I want to show chart with resources | 14:40 |
| lsmola | eglynn, that could work right? | 14:40 |
| eglynn | lsmola: it would take some non-trivial cross-querying of nova and ceilometer APIs to figure out which instances certain alarms apply to | 14:41 |
| lsmola | eglynn, it would be enough to look for alarms that have resource_id in their metadata, that would give alarm tied to that resource right? | 14:41 |
| eglynn | lsmola: yes, those map one-to-one to some individual resource | 14:41 |
| eglynn | lsmola: (which would be an instance where the target meter is instance-related) | 14:42 |
| lsmola | eglynn, then I could have different charts for projects, resource_aggregates, etc. | 14:42 |
| eglynn | lsmola: (or could be some other type of resource, e.g. an image or a volume, for other meter types ...) | 14:42 |
| *** SergeyLukjanov has quit IRC | 14:42 | |
| eglynn | lsmola: what do you mean by resource_aggregates? | 14:43 |
| lsmola | eglynn, yes, then I could just filter that by meter_names connected to nova, and I have nova instance health map, right? | 14:43 |
| lsmola | eglynn, sorry host aggregates etc | 14:43 |
| lsmola | eglynn, though host aggregates are not in ceilometer, so it degrades to list of resources, right? | 14:44 |
| eglynn | lsmola: I'm not sure that ceilo currently preserves the metadata identifying the host aggregate | 14:45 |
| lsmola | eglynn, no, I think not | 14:45 |
| eglynn | lsmola: yep, identifying alarms that refer one-to-one to instances should be possible | 14:45 |
| lsmola | eglynn, there is only resource, user, project | 14:45 |
| *** herndon has joined #openstack-metering | 14:46 | |
| lsmola | eglynn, ok should i write this into bp, or you will? | 14:46 |
| eglynn | lsmola: but non-trivial for alarms that refer to groups of instances | 14:46 |
| eglynn | lsmola: lets start with a discussion page on the wiki | 14:46 |
| eglynn | lsmola: can you capture your initial requirements there | 14:47 |
| eglynn | lsmola: then we can decide where we need to go from there, massage it into a BP etc. | 14:47 |
| eglynn | lsmola: make sense? | 14:49 |
| lsmola | eglynn, yes, but that query would target the group directly, e.g. health map or projects or users | 14:50 |
| lsmola | eglynn, well the horizon (and probably the tuskar) will probably need this | 14:51 |
| eglynn | lsmola: which query exactly do you mean there? ... "that query would target the group directly" | 14:51 |
| lsmola | eglynn, it's the same reason as for the-sample api, the chart would have to do to many queries to ceilometer and it's too slow, it needs to happen in one query | 14:51 |
| eglynn | lsmola: sure, I understand that's the desire | 14:52 |
| lsmola | eglynn, yeah like if the alarm is tied e.g. to user (and all of his resources) I could fetch those kind of alarms to show user based health map | 14:52 |
| lsmola | eglynn, it would probably need to add also all ids of the resources of the user, but that would still be around two queries | 14:54 |
| lsmola | eglynn, ok, wiki it is | 14:54 |
| eglynn | lsmola: yep, get your ideal requirements up on the wiki, then we can discuss what's acheivable | 14:55 |
| lsmola | eglynn, now how do I create a new wiki page? :-D | 14:55 |
| lsmola | eglynn, oh, it has separate sign in :-D | 14:56 |
| eglynn | lsmola: just point your broswer at https://wiki.openstack.org/wiki/Ceilometer/foobar | 14:56 |
| eglynn | lsmola: yeah, usual launchpad business to log in | 14:57 |
| lsmola | eglynn, oh cool | 14:59 |
| *** dhellmann_ is now known as dhellmann | 15:00 | |
| *** yanghy has quit IRC | 15:18 | |
| *** yanghy has joined #openstack-metering | 15:19 | |
| *** SergeyLukjanov has joined #openstack-metering | 15:21 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Use global openstack requirements https://review.openstack.org/46189 | 15:37 |
| lsmola | eglynn, I have filled some basic user stories https://wiki.openstack.org/wiki/Ceilometer/blueprints/alarm-audit-api-group-by#The_user_stories | 15:45 |
| lsmola | eglynn, is it enough to talk just about user stories, or i should make a proper draft of API, and the database implementation? | 15:46 |
| eglynn | lsmola: user stories are a good starting point, I'll review it later on today and get back to you with comments | 15:47 |
| lsmola | eglynn, ok cool, I will be going home in few minutes, I would think some more about the user stories tomorrow any feedback will be much appreciated | 15:51 |
| lsmola | eglynn, that you for your help :-) | 15:52 |
| lsmola | eglynn, that/thank :-D | 15:52 |
| eglynn | lsmola: cool ... BTW which TZ are you based in? | 15:52 |
| eglynn | lsmola: (I'm in GMT ...) | 15:52 |
| nealph | jd__: is there a utility to generate a message on the queue that the dispatchers will pick up? sandywalsh helped me find send-counter yesterday, but I believe it only pushes through the sample pipeline... | 15:53 |
| lsmola | eglynn, I am UTC +1 or 2 | 15:53 |
| lsmola | eglynn, Brno :-) | 15:53 |
| eglynn | lsmola: cool, that was my guess ;) | 15:53 |
| sandywalsh | apmelton1: dragondm do we have any utils for stuffing notifications on the queue that nealph might be able to use? | 15:54 |
| apmelton1 | sandywalsh: I never got around to cleaning this up, but it can be set up to throw notifications on rabbit | 15:55 |
| apmelton1 | https://github.com/ramielrowe/notif-gen | 15:55 |
| lsmola | eglynn, cool, thanks again for your help :-) see you tomorrow | 15:56 |
| nealph | sandywalsh: thanks! and btw, re: my question late yesterday on publisher output? see my bug https://bugs.launchpad.net/bugs/1224190 | 15:56 |
| nealph | apmelton1: sweet! thanks...I'll take a look. | 15:56 |
| nealph | sandywalsh: or rather, don't bother answering the question, because apparently it's a bug. :) | 15:57 |
| sandywalsh | nealph: <reading> | 15:57 |
| sandywalsh | nealph: yeah, definitely a bug. a str() missing :) | 16:00 |
| sandywalsh | nealph: sent you a side channel msg | 16:00 |
| *** Ruetobas has quit IRC | 16:01 | |
| *** boris-42_ has joined #openstack-metering | 16:02 | |
| *** Ruetobas has joined #openstack-metering | 16:04 | |
| *** boris-42_ has quit IRC | 16:04 | |
| *** yanghy has quit IRC | 16:05 | |
| *** boris-42 has quit IRC | 16:06 | |
| dragondm | sandywalsh: not at the moment :P | 16:07 |
| *** Ruetobas has quit IRC | 16:08 | |
| *** Ruetobas has joined #openstack-metering | 16:14 | |
| *** changbl has quit IRC | 16:14 | |
| *** eglynn has quit IRC | 16:30 | |
| openstackgerrit | litong01 proposed a change to openstack/ceilometer: refactor db2 get_meter_statistics method to support mongodb and db2 https://review.openstack.org/46175 | 16:39 |
| *** ruhe has joined #openstack-metering | 16:58 | |
| *** Bada has joined #openstack-metering | 17:05 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Update openstack.common.policy from oslo-incubator https://review.openstack.org/45210 | 17:07 |
| *** nati_ueno has joined #openstack-metering | 17:07 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Alarm history storage implementation for sqlalchemy https://review.openstack.org/45244 | 17:10 |
| *** nati_ueno has quit IRC | 17:12 | |
| *** ruhe has quit IRC | 17:13 | |
| thomasm | yay! | 17:41 |
| *** SergeyLukjanov has quit IRC | 17:42 | |
| *** thomasm has quit IRC | 17:42 | |
| Bada | Hi guys,i'm following the Manual installation doc of ceilometer. Where is it written to create a ceilometer user ? | 17:42 |
| *** Bada has quit IRC | 17:42 | |
| *** Bada has joined #openstack-metering | 17:43 | |
| *** SergeyLukjanov has joined #openstack-metering | 17:44 | |
| *** boris-42 has joined #openstack-metering | 17:53 | |
| *** nati_ueno has joined #openstack-metering | 18:00 | |
| *** eglynn has joined #openstack-metering | 18:00 | |
| *** thomasm has joined #openstack-metering | 18:04 | |
| *** herndon has quit IRC | 18:08 | |
| *** ruhe has joined #openstack-metering | 18:12 | |
| *** shaneduan has quit IRC | 18:29 | |
| *** shaneduan has joined #openstack-metering | 18:43 | |
| *** shaneduan is now known as Guest91769 | 18:43 | |
| *** Bada has quit IRC | 18:44 | |
| *** Bada has joined #openstack-metering | 18:45 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Add group by statistics examples in API v2 docs https://review.openstack.org/46181 | 18:46 |
| openstackgerrit | A change was merged to openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/46194 | 18:48 |
| *** jtran has joined #openstack-metering | 18:49 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Catch exceptions from nova client in poll_and_publish https://review.openstack.org/46203 | 18:51 |
| openstackgerrit | A change was merged to openstack/ceilometer: run-tests: fix MongoDB start wait https://review.openstack.org/45642 | 18:51 |
| *** Guest91769 is now known as shaneduan | 18:56 | |
| *** lnxnut_ has joined #openstack-metering | 18:57 | |
| *** changbl has joined #openstack-metering | 19:05 | |
| *** herndon has joined #openstack-metering | 19:06 | |
| *** ruhe has quit IRC | 19:08 | |
| *** toddmck has joined #openstack-metering | 19:13 | |
| *** gordc has quit IRC | 19:32 | |
| *** zul has quit IRC | 19:45 | |
| *** SergeyLukjanov has quit IRC | 19:54 | |
| *** zul has joined #openstack-metering | 19:59 | |
| *** sandywalsh has quit IRC | 20:20 | |
| *** lnxnut_ has quit IRC | 20:30 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Improve libvirt vnic parsing with missing mac! https://review.openstack.org/45446 | 20:38 |
| *** Bada has quit IRC | 21:07 | |
| *** rrao is now known as rrao-away | 21:15 | |
| *** rrao-away is now known as rrao_away | 21:17 | |
| *** rrao_away is now known as zz_rrao | 21:18 | |
| *** zz_rrao is now known as rrao | 21:18 | |
| *** shang has quit IRC | 21:38 | |
| *** toddmck has left #openstack-metering | 21:49 | |
| *** eglynn has quit IRC | 21:49 | |
| *** d34dh0r53 has joined #openstack-metering | 21:49 | |
| *** d34dh0r53 has quit IRC | 21:51 | |
| *** thomasm has quit IRC | 21:53 | |
| *** boris-42 has quit IRC | 21:57 | |
| *** changbl has quit IRC | 22:00 | |
| *** bpokorny has quit IRC | 22:04 | |
| *** shang has joined #openstack-metering | 23:02 | |
| *** zaneb has joined #openstack-metering | 23:12 | |
| *** zbitter has quit IRC | 23:14 | |
| *** herndon has quit IRC | 23:22 | |
| openstackgerrit | Nachi Ueno proposed a change to openstack/ceilometer: Add ElasticSearch Support for ceilometer https://review.openstack.org/46383 | 23:28 |
| openstackgerrit | Nachi Ueno proposed a change to openstack/ceilometer: Add ElasticSearch Support for ceilometer https://review.openstack.org/46383 | 23:35 |
| *** shang has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!