| *** flwang has quit IRC | 00:07 | |
| *** flwang1 has joined #openstack-telemetry | 00:09 | |
| *** swamireddy has quit IRC | 00:20 | |
| *** swamireddy has joined #openstack-telemetry | 00:23 | |
| *** flwang1 has quit IRC | 00:23 | |
| *** flwang1 has joined #openstack-telemetry | 00:26 | |
| *** vishwanathj has joined #openstack-telemetry | 00:34 | |
| *** jwcroppe has quit IRC | 00:34 | |
| *** jwcroppe has joined #openstack-telemetry | 00:35 | |
| *** thorst has joined #openstack-telemetry | 00:37 | |
| *** jwcroppe has quit IRC | 00:39 | |
| *** jwcroppe has joined #openstack-telemetry | 00:42 | |
| marcin12345 | gordc: what is the exploding bug about? | 00:43 |
|---|---|---|
| marcin12345 | that worked for me, but I had to comment line 76 in /usr/lib/python2.7/site-packages/gnocchi/storage/swift.py | 00:44 |
| *** chlong has joined #openstack-telemetry | 00:54 | |
| *** jwcroppe has quit IRC | 00:55 | |
| *** diogogmt has joined #openstack-telemetry | 00:56 | |
| *** jwcroppe has joined #openstack-telemetry | 01:06 | |
| *** jwcroppe has quit IRC | 01:14 | |
| *** ljxiash has quit IRC | 01:15 | |
| *** chlong has quit IRC | 01:18 | |
| marcin12345 | gordc: is there a way to specify storage policy for Gnocchi so it does not use default policy? The way it is now it creates a lot of containers in the same space that users are, which is default :( | 01:20 |
| *** chlong has joined #openstack-telemetry | 01:30 | |
| *** caisan has joined #openstack-telemetry | 01:38 | |
| *** ljxiash has joined #openstack-telemetry | 01:41 | |
| *** mdih has joined #openstack-telemetry | 01:44 | |
| *** ljxiash has quit IRC | 01:55 | |
| openstackgerrit | liusheng proposed openstack/python-aodhclient: Add --filter to "alarm list" https://review.openstack.org/292233 | 02:04 |
| *** ljxiash has joined #openstack-telemetry | 02:05 | |
| openstackgerrit | liusheng proposed openstack/python-aodhclient: Add --filter to "alarm list" https://review.openstack.org/292233 | 02:07 |
| *** purplerbot has quit IRC | 02:08 | |
| *** purplerbot has joined #openstack-telemetry | 02:09 | |
| *** thorst has quit IRC | 02:09 | |
| *** thorst has joined #openstack-telemetry | 02:10 | |
| *** thorst has quit IRC | 02:18 | |
| *** dave-mccowan has quit IRC | 02:19 | |
| *** kbyrne has quit IRC | 02:19 | |
| *** sileht has quit IRC | 02:19 | |
| *** llu has quit IRC | 02:19 | |
| *** fc__ has quit IRC | 02:19 | |
| *** dave-mccowan has joined #openstack-telemetry | 02:25 | |
| *** kbyrne has joined #openstack-telemetry | 02:25 | |
| *** sileht has joined #openstack-telemetry | 02:25 | |
| *** llu has joined #openstack-telemetry | 02:25 | |
| *** fc__ has joined #openstack-telemetry | 02:25 | |
| *** kbyrne has quit IRC | 02:25 | |
| openstackgerrit | Merged openstack/python-aodhclient: Use overtest to run Gnocchi in tests https://review.openstack.org/292515 | 02:26 |
| *** kbyrne has joined #openstack-telemetry | 02:28 | |
| *** ljxiash has quit IRC | 02:31 | |
| openstackgerrit | Merged openstack/python-aodhclient: tox: remove useless install_command https://review.openstack.org/292517 | 02:41 |
| openstackgerrit | Merged openstack/python-aodhclient: Clean flake8 paths and unused config options https://review.openstack.org/292791 | 02:41 |
| *** ljxiash has joined #openstack-telemetry | 02:42 | |
| *** liamji has joined #openstack-telemetry | 02:43 | |
| *** slogan has quit IRC | 03:15 | |
| *** thorst has joined #openstack-telemetry | 03:16 | |
| *** jwcroppe has joined #openstack-telemetry | 03:19 | |
| *** thorst has quit IRC | 03:23 | |
| *** links has joined #openstack-telemetry | 03:45 | |
| openstackgerrit | liusheng proposed openstack/gnocchi: Fix gate: add uuid-runtime package installation https://review.openstack.org/293808 | 03:47 |
| *** dave-mccowan has quit IRC | 04:00 | |
| *** ljxiash has quit IRC | 04:01 | |
| *** rbak has joined #openstack-telemetry | 04:02 | |
| *** jwcroppe has quit IRC | 04:19 | |
| *** jwcroppe has joined #openstack-telemetry | 04:19 | |
| *** thorst has joined #openstack-telemetry | 04:21 | |
| *** rbak has quit IRC | 04:21 | |
| *** jwcroppe has quit IRC | 04:23 | |
| *** thorst has quit IRC | 04:28 | |
| *** ljxiash has joined #openstack-telemetry | 04:43 | |
| *** ljxiash has quit IRC | 05:00 | |
| *** thorst has joined #openstack-telemetry | 05:26 | |
| *** _nadya_ has joined #openstack-telemetry | 05:27 | |
| *** thorst has quit IRC | 05:33 | |
| *** ljxiash has joined #openstack-telemetry | 05:35 | |
| *** chlong has quit IRC | 06:08 | |
| *** _nadya_ has quit IRC | 06:10 | |
| *** vishwanathj is now known as vishwanathj_zzz | 06:12 | |
| openstackgerrit | liusheng proposed openstack/python-aodhclient: Correct and add examples in docs of CLI https://review.openstack.org/293835 | 06:13 |
| *** chlong has joined #openstack-telemetry | 06:20 | |
| *** thorst has joined #openstack-telemetry | 06:31 | |
| *** BigWillie has joined #openstack-telemetry | 06:33 | |
| *** BigWillie has quit IRC | 06:34 | |
| *** thorst has quit IRC | 06:38 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/aodh: devstack: allow uwsgi deployments https://review.openstack.org/292245 | 06:42 |
| *** _nadya_ has joined #openstack-telemetry | 07:01 | |
| *** yprokule has joined #openstack-telemetry | 07:08 | |
| *** tesseract has joined #openstack-telemetry | 07:15 | |
| *** tesseract is now known as Guest57182 | 07:15 | |
| *** shakamunyi has quit IRC | 07:18 | |
| *** _nadya_ has quit IRC | 07:18 | |
| *** shakamunyi has joined #openstack-telemetry | 07:24 | |
| *** _nadya_ has joined #openstack-telemetry | 07:32 | |
| *** thorst has joined #openstack-telemetry | 07:36 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/aodh: Raise Error when query history of an alarm that are not existed https://review.openstack.org/293861 | 07:37 |
| *** rcernin has joined #openstack-telemetry | 07:39 | |
| *** thorst has quit IRC | 07:43 | |
| *** belmoreira has joined #openstack-telemetry | 07:52 | |
| *** safchain has joined #openstack-telemetry | 07:54 | |
| *** chlong has quit IRC | 07:57 | |
| *** chlong has joined #openstack-telemetry | 07:58 | |
| *** chlong has quit IRC | 08:02 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/aodh: Raise Error when query history of an alarm that are not existed https://review.openstack.org/293861 | 08:21 |
| *** pcaruana has joined #openstack-telemetry | 08:24 | |
| *** thorst has joined #openstack-telemetry | 08:40 | |
| *** barra204 has joined #openstack-telemetry | 08:43 | |
| *** shakamunyi has quit IRC | 08:44 | |
| *** thorst has quit IRC | 08:48 | |
| *** links has quit IRC | 08:53 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/aodh: Raise Error when query history of an alarm that are not existed https://review.openstack.org/293861 | 08:53 |
| *** yassine__ has joined #openstack-telemetry | 09:25 | |
| openstackgerrit | Julien Danjou proposed openstack/aodh: api: stop relying on side-effect of _alarm() https://review.openstack.org/293902 | 09:29 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: api: rename _alarm to _enforce_rbac https://review.openstack.org/293903 | 09:29 |
| *** chlong has joined #openstack-telemetry | 09:31 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ceilometer: Imported Translations from Zanata https://review.openstack.org/292715 | 09:35 |
| *** thorst has joined #openstack-telemetry | 09:45 | |
| *** ljxiash has quit IRC | 09:52 | |
| *** thorst has quit IRC | 09:52 | |
| *** _nadya_ has quit IRC | 10:01 | |
| *** nadya has joined #openstack-telemetry | 10:04 | |
| *** ljxiash has joined #openstack-telemetry | 10:06 | |
| openstackgerrit | Lianhao Lu proposed openstack/gnocchi: Added docs about new snmp related resource types https://review.openstack.org/293917 | 10:06 |
| *** ljxiash has quit IRC | 10:10 | |
| *** fwdit has joined #openstack-telemetry | 10:11 | |
| *** caisan has quit IRC | 10:12 | |
| *** cdent has joined #openstack-telemetry | 10:22 | |
| *** links has joined #openstack-telemetry | 10:29 | |
| *** albus has joined #openstack-telemetry | 10:32 | |
| albus | Hey Guys, I just want to know how i can query the alarms based on severity via the aodh client | 10:33 |
| albus | Any input would be of great help | 10:33 |
| *** chlong has quit IRC | 10:39 | |
| *** thorst has joined #openstack-telemetry | 10:51 | |
| *** thorst has quit IRC | 10:58 | |
| *** ljxiash has joined #openstack-telemetry | 11:03 | |
| *** yassine__ has quit IRC | 11:04 | |
| *** boris-42 has joined #openstack-telemetry | 11:14 | |
| nadya | @cdent Hi Chris! Probably, you remember what's the final decision about aodhclient? In Mitaka ceilometer client is still recommended, right? | 11:15 |
| cdent | nadya: sorry I don't remember what the final decision was | 11:16 |
| nadya | cdent: yep... understood. Will try to check the logs | 11:16 |
| nadya | liusheng: ^^ hi! | 11:16 |
| *** liamji has quit IRC | 11:17 | |
| nadya | yep, so based on the previous meeting, looks like we still recommend ceilometerclient for aodh in Mitaka | 11:21 |
| *** albus has quit IRC | 11:35 | |
| *** cdent has quit IRC | 11:37 | |
| *** thorst has joined #openstack-telemetry | 11:42 | |
| *** thorst has quit IRC | 11:43 | |
| *** thorst has joined #openstack-telemetry | 11:43 | |
| *** thorst_ has joined #openstack-telemetry | 11:44 | |
| *** chlong has joined #openstack-telemetry | 11:48 | |
| *** thorst has quit IRC | 11:48 | |
| *** openstackgerrit has quit IRC | 11:48 | |
| *** openstackgerrit has joined #openstack-telemetry | 11:49 | |
| *** dave-mccowan has joined #openstack-telemetry | 11:53 | |
| *** yassine__ has joined #openstack-telemetry | 11:59 | |
| *** cdent has joined #openstack-telemetry | 12:03 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/aodh: Raise Error when query history of an alarm that are not existed https://review.openstack.org/293861 | 12:11 |
| *** thorst has joined #openstack-telemetry | 12:17 | |
| *** gordc has joined #openstack-telemetry | 12:18 | |
| *** thorst_ has quit IRC | 12:21 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/aodh: Raise Error when query history of an alarm that are not existed https://review.openstack.org/293861 | 12:39 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Test foobar https://review.openstack.org/293990 | 12:40 |
| *** dikonoor has joined #openstack-telemetry | 12:44 | |
| *** dikonoor has quit IRC | 12:46 | |
| *** dikonoor has joined #openstack-telemetry | 12:46 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Test foobar https://review.openstack.org/293990 | 12:47 |
| *** dikonoor has quit IRC | 12:47 | |
| gordc | jd__: should this be part of release candidate? https://review.openstack.org/#/c/293902/ | 12:53 |
| jd__ | gordc: no it's just a cleanup patch | 12:53 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Test foobar https://review.openstack.org/293990 | 12:53 |
| *** diogogmt has joined #openstack-telemetry | 12:55 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Test foobar https://review.openstack.org/293990 | 12:55 |
| gordc | jd__: cool cool. | 12:56 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Test foobar https://review.openstack.org/293990 | 12:56 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: api: rename _alarm to _enforce_rbac https://review.openstack.org/293903 | 12:57 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: api: stop relying on side-effect of _alarm() https://review.openstack.org/293902 | 12:57 |
| *** jwcroppe has joined #openstack-telemetry | 12:59 | |
| *** peristeri has joined #openstack-telemetry | 12:59 | |
| gordc | sileht: i'm going to leave https://review.openstack.org/#/c/292245/ off mitaka unless you prefer it in. i can't cut mitaka anyways until gate is fixed. | 13:00 |
| sileht | gordc, no problem | 13:16 |
| *** pradk has joined #openstack-telemetry | 13:18 | |
| gordc | sileht: cool cool. let me know if you change your mind. i don't think this gate thing will work for a while | 13:18 |
| *** diogogmt has quit IRC | 13:18 | |
| sileht | gordc, ok, we need to wait next image build of something ? | 13:18 |
| gordc | sileht: i think so. the integration fails because uuidgen doesn't exist on new trusty images | 13:25 |
| sileht | gordc, your fix have been merged, no ? | 13:25 |
| gordc | i tried to install with devstack: https://review.openstack.org/#/c/293735/ | 13:25 |
| gordc | but apparently it doesn't, install until after | 13:26 |
| gordc | i see this patch now: https://review.openstack.org/#/c/293942/ | 13:26 |
| *** rbak has joined #openstack-telemetry | 13:26 | |
| gordc | so basically my patch and all the backports probably don't do anything | 13:26 |
| *** links has quit IRC | 13:28 | |
| sileht | :( | 13:29 |
| gordc | yeah. i didn't think my patch would work. i'm usually just guessing when i patch gate stuff | 13:30 |
| sileht | gordc, the funny part is that review seems agreed on your change :) | 13:33 |
| sileht | review/reviewer | 13:33 |
| gordc | sileht: lol that's because i put 'fix' in title. of course it fixes nothing | 13:35 |
| openstackgerrit | Zhenyu Zheng proposed openstack/aodh: Raise Error when query history of an alarm that are not existed https://review.openstack.org/293861 | 13:39 |
| *** shardy has joined #openstack-telemetry | 13:41 | |
| *** cdent has quit IRC | 13:41 | |
| *** rbak has quit IRC | 13:47 | |
| *** KrishR has joined #openstack-telemetry | 14:00 | |
| *** liamji has joined #openstack-telemetry | 14:02 | |
| *** rbak has joined #openstack-telemetry | 14:11 | |
| *** nicodemus_ has joined #openstack-telemetry | 14:21 | |
| *** pcaruana has quit IRC | 14:28 | |
| *** tongli has joined #openstack-telemetry | 14:29 | |
| *** mriedem has joined #openstack-telemetry | 14:47 | |
| *** mriedem has left #openstack-telemetry | 14:49 | |
| *** belmoreira has quit IRC | 14:59 | |
| nicodemus_ | morning gordc | 14:59 |
| *** nadya has quit IRC | 14:59 | |
| *** ljxiash has quit IRC | 14:59 | |
| *** nadya has joined #openstack-telemetry | 14:59 | |
| *** cdent has joined #openstack-telemetry | 15:00 | |
| gordc | nicodemus_: hey, we're having our weekly meeting right now over at openstack-meeting | 15:00 |
| gordc | will need a few minutes. you welcome to join us | 15:01 |
| gordc | sileht: jd__: llu: pradk: ildikov: liusheng: meeting | 15:01 |
| nicodemus_ | gordc, ok! I'm joining | 15:02 |
| *** nadya has quit IRC | 15:02 | |
| gordc | nadya: ping | 15:02 |
| *** ityaptin has joined #openstack-telemetry | 15:02 | |
| gordc | r-mibu: courtesy ping | 15:02 |
| *** nadya has joined #openstack-telemetry | 15:02 | |
| *** ljxiash has joined #openstack-telemetry | 15:02 | |
| *** zqfan has joined #openstack-telemetry | 15:03 | |
| *** rbak has quit IRC | 15:03 | |
| *** rbak has joined #openstack-telemetry | 15:04 | |
| *** alejandrito has joined #openstack-telemetry | 15:05 | |
| *** vishwanathj_zzz is now known as vishwanathj | 15:17 | |
| gordc | nicodemus_: whatsup? | 15:18 |
| *** liusheng has quit IRC | 15:18 | |
| nicodemus_ | gordc, one question regarding gnocchi aggregation workers | 15:19 |
| *** liusheng has joined #openstack-telemetry | 15:19 | |
| gordc | sure | 15:19 |
| nicodemus_ | I've noticed that, as I increase the number of gnocchi workers the number of gnocchi-metricd processes also increases | 15:19 |
| nicodemus_ | however, as I increase the number of aggregation_workers in gnocchi.conf, the number of gnocchi-metricd workers stays the same | 15:20 |
| nicodemus_ | Is that correct? | 15:20 |
| gordc | nicodemus_: yep. the metricd workers actually spawn additional processes... aggregation_workers just spawns a thread to handle task of calculating the aggregate values in your policy file as they are all stored/calculated separately (sum,mean,max,min, etc...) | 15:22 |
| *** fwdit has quit IRC | 15:23 | |
| nicodemus_ | gordc, great! Thanks | 15:25 |
| gordc | yeah, so basically aggregation_workers is subtask within metricd. | 15:25 |
| gordc | nicodemus_: you still seeing the slow down in processing? | 15:26 |
| nicodemus_ | FYI, we've updated the bug report regarding the 'measure' object in ceph (#1557724) | 15:26 |
| nicodemus_ | 24 metricd workers + 2 aggregation workers vs 24 metricd workers + 8 aggregation workers, I didn't see much difference unfortunately | 15:27 |
| gordc | nicodemus_: cool cool. i think sileht is playing around with a solution | 15:27 |
| gordc | https://review.openstack.org/#/c/293990/ | 15:27 |
| gordc | nicodemus_: we may need to bring in something like redis. we just ideally don't want to bring in dependencies until we're absolutely sure | 15:28 |
| sileht | that doesn't work for now, and dealling with ceph lock is horrible, they are not automatically release when the connection is lost | 15:28 |
| nicodemus_ | gordc, that's great news! | 15:28 |
| gordc | sileht: that sucks. | 15:28 |
| nicodemus_ | In the meantime I'm doing my best to reduce the number of metrics collected | 15:29 |
| gordc | nicodemus_: you probably can but tbh, i don't think you have a lot of metrics. | 15:30 |
| nicodemus_ | gordc, I've noticed that the resource type instance and the resource type instance_disk have the same metrics for the root disk | 15:31 |
| nicodemus_ | plus there are a few that we don't really use, such as network.incoming/outgoing.bytes/packets | 15:32 |
| nicodemus_ | those we use .rate and .delta | 15:33 |
| gordc | they should be different measures (instance vs instance_disk). the latter should be per device metrics | 15:33 |
| gordc | yeah, you can/should filter out metrics in pipeline that you don't ned | 15:33 |
| gordc | need* | 15:33 |
| *** ljxiash has quit IRC | 15:43 | |
| *** nadya has quit IRC | 15:53 | |
| *** dhague_ has joined #openstack-telemetry | 16:02 | |
| *** diogogmt has joined #openstack-telemetry | 16:04 | |
| *** pcaruana has joined #openstack-telemetry | 16:22 | |
| *** jwcroppe has quit IRC | 16:22 | |
| *** jwcroppe has joined #openstack-telemetry | 16:23 | |
| *** mdih has quit IRC | 16:25 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Test foobar https://review.openstack.org/293990 | 16:25 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: devstack: allow uwsgi deployments https://review.openstack.org/292077 | 16:25 |
| *** yprokule has quit IRC | 16:39 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Test foobar https://review.openstack.org/293990 | 16:41 |
| *** yarkot_ has joined #openstack-telemetry | 16:48 | |
| *** alejandrito has quit IRC | 16:49 | |
| *** alejandrito has joined #openstack-telemetry | 16:51 | |
| *** nadya has joined #openstack-telemetry | 16:52 | |
| *** yassine__ has quit IRC | 16:53 | |
| *** nadya has quit IRC | 16:53 | |
| *** alejandrito has quit IRC | 17:01 | |
| *** nadya has joined #openstack-telemetry | 17:18 | |
| *** yarkot_ has quit IRC | 17:19 | |
| *** dhague_ has quit IRC | 17:19 | |
| *** cdent has quit IRC | 17:21 | |
| *** nadya has quit IRC | 17:22 | |
| *** cdent has joined #openstack-telemetry | 17:23 | |
| *** cdent has quit IRC | 17:25 | |
| *** cdent has joined #openstack-telemetry | 17:33 | |
| *** Guest57182 has quit IRC | 17:36 | |
| *** cdent has quit IRC | 17:44 | |
| nicodemus_ | gordc, can I ask you another question? Would you recommend trying gnocchi with influxdb for storage instead of ceph? | 17:49 |
| *** nadya has joined #openstack-telemetry | 17:50 | |
| *** vishwana_ has joined #openstack-telemetry | 17:51 | |
| gordc | nicodemus_: nope. not because i think influxdb sucks but because we stopped working on it a while ago | 17:52 |
| gordc | nicodemus_: i know we had a lot of issues and opened some blockers against influxdb. we just continued with our own implementation because we didn't have a viable influxdb solution | 17:53 |
| gordc | nicodemus_: long story short, don't try influxdb driver becuase it's not working | 17:53 |
| *** vishwanathj has quit IRC | 17:54 | |
| *** rcernin has quit IRC | 17:55 | |
| nicodemus_ | gordc, I see. It appears that so many workers impacted on VMs performance | 17:59 |
| gordc | nicodemus_: this is with 24 workers on two 24core machines, right? | 18:02 |
| nicodemus_ | yes, a total of 48 workers | 18:02 |
| gordc | ok. i'd hope that be more than enough given input. i'm going to see if there's anything i can fix with parallelism. | 18:07 |
| gordc | will be back in a bit. | 18:07 |
| *** gordc has quit IRC | 18:07 | |
| *** nadya has quit IRC | 18:15 | |
| *** liamji has quit IRC | 18:17 | |
| *** pcaruana has quit IRC | 18:18 | |
| *** pcaruana has joined #openstack-telemetry | 18:30 | |
| *** cdent has joined #openstack-telemetry | 18:31 | |
| *** KrishR has quit IRC | 18:38 | |
| *** iberezovskiy is now known as iberezovskiy|afk | 18:38 | |
| *** KrishR has joined #openstack-telemetry | 18:40 | |
| *** tongli has quit IRC | 18:41 | |
| *** gordc has joined #openstack-telemetry | 18:44 | |
| stevelle | gordc: are you suggesting tooz or redis specifically? | 18:45 |
| gordc | stevelle: for what? | 18:49 |
| stevelle | gordc: for that ceph locking bit | 18:50 |
| stevelle | that was 3.5h ago | 18:50 |
| gordc | stevelle: oh.lol let me take a look at chat history | 18:51 |
| gordc | oh. i was referring to how we store our unprocessed measures on the same backend as whatever driver you selected to stores data. so processed and unprocessed data sit on either ceph|swift|file... wwhere as we could technically store the unprocessed data in some intermediary storage such as redis (if that's deemed the limiting factor) | 18:53 |
| stevelle | gordc: ok, so I guess I'm wondering about the intermediary storage. Are you imagining that possibly being satisfied by either a storage driver or a tooz based driver? | 18:57 |
| gordc | well tooz doesn't actually store anything aside from group members | 18:58 |
| *** jmlowe has joined #openstack-telemetry | 19:00 | |
| *** pradk has quit IRC | 19:02 | |
| gordc | i think it does relate to either how we're storing the unprocessed data or how we're parallelising it. for me, i don't really see any diffrence in performance between a metric with 1 unprocessed datapoint and a metric with 500 unprocessed datapoints | 19:03 |
| stevelle | I don't recall but I looked at tooz, saw there were things besides locks that you could use it for. | 19:10 |
| stevelle | I just wanted to make sure I was understanding the possible impacts on deployments if a new tech was being laid out, and what the HA story would look like since redis doesn't have a great HA story. | 19:11 |
| stevelle | will think on it more | 19:12 |
| jmlowe | I seem to be having trouble getting notification stuff in gnocchi, I haven't seen any obvious errors | 19:12 |
| gordc | yeah, that's a good point. it's probably something of last resort. in theory, we could also just shove unprocessed measures on a queue. | 19:13 |
| gordc | jmlowe: what specific stuff? | 19:13 |
| jmlowe | I get image,instance,instance_disk,instance_network_interface but I'm expecting at least some keystone, cinder, ipmi resource types to show up | 19:13 |
| gordc | do you have them listed in gnocchi_resources.yaml? | 19:15 |
| gordc | jmlowe: the keystone stuff won't make any sense in gnocchi since it doesn't actually contain any measuremetns | 19:15 |
| *** pradk has joined #openstack-telemetry | 19:17 | |
| jmlowe | ah, ok, it's filtered as being one of the blacklisted types since gnocchi_url is set? | 19:18 |
| jmlowe | how about ipmi? | 19:19 |
| gordc | well the keystone stuff will be ignored if disable_non_metric_meters is set. | 19:20 |
| gordc | the ipmi stuff should be in gnocchi assuming you are actually polling htem | 19:20 |
| jmlowe | I was naively assuming you could just start one per hypervisor a la compute and it would read the local counters from /sys, is that not the case? | 19:22 |
| gordc | you need an ipmi agent | 19:23 |
| gordc | http://docs.openstack.org/admin-guide-cloud/telemetry-data-collection.html#ipmi-agent you'll need to configure it explicitly in pipeline if you don't use ironic | 19:24 |
| openstackgerrit | Atul Aggarwal proposed openstack/ceilometer: Enhancing Retry logic to Coordination when joining partitioning grp https://review.openstack.org/294244 | 19:24 |
| *** shardy has quit IRC | 19:36 | |
| *** davidlenwell has quit IRC | 19:36 | |
| *** nadya has joined #openstack-telemetry | 19:39 | |
| *** davidlenwell has joined #openstack-telemetry | 19:42 | |
| jmlowe | gordc: I'm confused by the identity types in /etc/ceilometer/gnocchi_resources.yaml | 19:46 |
| gordc | jmlowe: i think most people are. i would argue they shouldn't be there since the measurement value is always = '1'. they are basically healthcheck events so they aren't really measuring anything. just confirming something exists | 19:49 |
| *** pradk has quit IRC | 20:08 | |
| *** aaggarw3 has joined #openstack-telemetry | 20:09 | |
| openstackgerrit | Atul Aggarwal proposed openstack/ceilometer: Enhancing Retry logic to Coordination when joining partitioning grp https://review.openstack.org/294244 | 20:16 |
| *** pcaruana has quit IRC | 20:17 | |
| jmlowe | what am I missing that would cause ERROR ceilometer.agent.manager [-] Skip loading extension for hardware.ipmi.node.power ? | 20:22 |
| *** vishwanathj has joined #openstack-telemetry | 20:23 | |
| gordc | nicodemus_: can you try setting hte aggregation_workers to 1? | 20:24 |
| nicodemus_ | gordc, yes | 20:24 |
| nicodemus_ | what should I expect? | 20:24 |
| *** vishwana_ has quit IRC | 20:25 | |
| gordc | nicodemus_: i'm not sure. i think the aggregation_workers does nothing. or actually performs worse. | 20:26 |
| gordc | i think it's someting related to stupid python GIL | 20:26 |
| gordc | when i try locally, it's actually better if i just make it single-threaded | 20:27 |
| gordc | jmlowe: http://lists.openstack.org/pipermail/openstack-dev/2016-March/088314.html | 20:27 |
| gordc | nicodemus_: disclaimer: "not much better, but it's not worse." | 20:28 |
| nicodemus_ | gordc, hahaha got it | 20:29 |
| *** pcaruana has joined #openstack-telemetry | 20:30 | |
| *** rbak has quit IRC | 20:40 | |
| *** ametts has joined #openstack-telemetry | 20:42 | |
| *** cdent has quit IRC | 20:47 | |
| *** cdent has joined #openstack-telemetry | 20:47 | |
| nicodemus_ | gordc, I couldn't appreciate much difference with agregation_workers = 1 | 20:48 |
| nicodemus_ | perhaps after some time the trend might become visible | 20:49 |
| gordc | cool cool. yeah, that's what i figured. i just want to confirm that the threading part doesn't actually parallelise anything. | 20:50 |
| *** rbak has joined #openstack-telemetry | 20:52 | |
| nicodemus_ | gordc, let me ask you another question, just to be sure that I'm editing the pipeline.yaml file correctly | 20:58 |
| gordc | sure | 21:00 |
| nicodemus_ | for example, I have a cpu_source that has "cpu" as its only meter, and the "instance" resource in gnocchi_resources.yaml has cpu, cpu_util and cpu.delta between its metrics | 21:00 |
| nicodemus_ | if I want to have only cpu_util and cpu.delta, should I explicitly add those two meters in the cpu_source instead of just "cpu"? | 21:01 |
| nicodemus_ | (let me know if I explained myself) | 21:01 |
| gordc | what does your pipeline.yaml look like? | 21:03 |
| gordc | basically, the meters you define in pipeline.yaml are what get genreated | 21:03 |
| nicodemus_ | Here it is: http://paste.openstack.org/show/490990/ | 21:03 |
| *** cdent has quit IRC | 21:04 | |
| gordc | if all you want is cpu_util and cpu.delta, it would look like http://paste.openstack.org/show/490993/ | 21:05 |
| gordc | which basically just grabs cpu time and forces it through two sinks, one sink which converts it into delta another which converts it into utilisation | 21:05 |
| nicodemus_ | I see, and then where the ceilometer-collector service is running, I choose exactly what metrics each gnocchi resoure should have | 21:08 |
| nicodemus_ | removing the ones I don't want/use | 21:08 |
| openstackgerrit | Atul Aggarwal proposed openstack/ceilometer: Enhancing Retry logic to Coordination when joining partitioning grp https://review.openstack.org/294244 | 21:09 |
| *** cdent has joined #openstack-telemetry | 21:10 | |
| gordc | nicodemus_: no, it's already been filtered by the polling/notificaiton agents | 21:11 |
| gordc | the gnocchi_resources.yaml file is just a mapping on how to take ceilometer meter and input into gnocchi | 21:12 |
| *** cdent has quit IRC | 21:13 | |
| *** cdent has joined #openstack-telemetry | 21:13 | |
| nicodemus_ | I believe I wasn't 100% clear, I don't want *only* cpu_util and cpu.delta, I still want to have instance, disk, and so on | 21:14 |
| nicodemus_ | but regading the cpu_source, I want cpu_util and cpu.delta instead of cpu_util, cpu.delta and cpu | 21:14 |
| *** nadya has quit IRC | 21:15 | |
| gordc | oh. the cpu meter isn't coming from that cpu_source. it's actually coming from meter_source | 21:15 |
| gordc | that basically is building/emitting everything except for the instance meter. | 21:16 |
| gordc | you should add !cpu to that if you don't want cpu at all | 21:16 |
| *** cdent has quit IRC | 21:16 | |
| jmlowe | ok, I give up on ipmi for now | 21:16 |
| nicodemus_ | got it! | 21:16 |
| nicodemus_ | thanks a lot :) | 21:17 |
| gordc | np | 21:17 |
| gordc | nicodemus_: just an fyi, cloudkitty doesn't have a ptl nomination (it doesn't really matter if everything stays the same) | 21:18 |
| gordc | http://lists.openstack.org/pipermail/openstack-dev/2016-March/089708.html | 21:19 |
| nicodemus_ | gordc, what does it mean? | 21:21 |
| nicodemus_ | or more specifically, what do you mean? | 21:22 |
| gordc | nicodemus_: big tent projects are required to have PTLs nominations/elections. it seems cloudkitty project doesn't have one. | 21:24 |
| gordc | just figured i'd mentinon it. i need to drop off. ttyl | 21:25 |
| *** gordc has quit IRC | 21:25 | |
| *** diogogmt has quit IRC | 21:30 | |
| *** cdent has joined #openstack-telemetry | 21:31 | |
| *** cdent has quit IRC | 21:33 | |
| *** thorst has quit IRC | 21:42 | |
| *** thorst has joined #openstack-telemetry | 21:43 | |
| *** nicodemus_ has quit IRC | 21:43 | |
| *** thorst has quit IRC | 21:47 | |
| *** zqfan has quit IRC | 22:02 | |
| *** peristeri has quit IRC | 22:05 | |
| *** cdent has joined #openstack-telemetry | 22:05 | |
| marcin12345 | hey gordc: do you know if there is a way to specify storage policy for Swift when using it by Gnocchi? | 22:07 |
| *** ametts has quit IRC | 22:14 | |
| *** cdent has quit IRC | 22:19 | |
| *** r-mibu has quit IRC | 22:20 | |
| *** thorst has joined #openstack-telemetry | 22:21 | |
| *** r-mibu has joined #openstack-telemetry | 22:30 | |
| *** KrishR has quit IRC | 22:35 | |
| *** diogogmt has joined #openstack-telemetry | 22:51 | |
| *** alextricity25 has left #openstack-telemetry | 23:00 | |
| *** tristanC has joined #openstack-telemetry | 23:15 | |
| *** rbak has quit IRC | 23:21 | |
| *** gzhai has quit IRC | 23:43 | |
| *** gzhai has joined #openstack-telemetry | 23:47 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!