| *** gmann has quit IRC | 01:43 | |
| *** zhurong has quit IRC | 02:35 | |
| *** altlogbot_3 has quit IRC | 02:41 | |
| *** altlogbot_1 has joined #openstack-watcher | 02:44 | |
| *** altlogbot_1 has quit IRC | 03:31 | |
| *** altlogbot_0 has joined #openstack-watcher | 03:33 | |
| *** zhurong has joined #openstack-watcher | 05:24 | |
| openstackgerrit | chenker proposed openstack/watcher master: Put the method add_migration() in base.py https://review.opendev.org/655675 | 05:55 |
|---|---|---|
| openstackgerrit | chenker proposed openstack/watcher master: Put the method add_migration() in base.py https://review.opendev.org/655675 | 05:57 |
| openstackgerrit | chenker proposed openstack/watcher master: Put the method add_migration() in base.py https://review.opendev.org/655675 | 06:10 |
| *** adisky__ has joined #openstack-watcher | 06:16 | |
| openstackgerrit | chenker proposed openstack/watcher master: Put the method add_migration() in base.py https://review.opendev.org/655675 | 06:29 |
| openstackgerrit | Merged openstack/watcher-specs master: Make spec's line limited to a maximum of 79 characters https://review.opendev.org/657184 | 06:54 |
| openstackgerrit | Merged openstack/watcher-specs master: update git.openstack.org to opendev https://review.opendev.org/656246 | 06:54 |
| *** altlogbot_0 has quit IRC | 07:13 | |
| openstackgerrit | chenker proposed openstack/watcher-specs master: Use pep8 replace py27 https://review.opendev.org/657192 | 07:16 |
| *** altlogbot_1 has joined #openstack-watcher | 07:19 | |
| *** altlogbot_1 has quit IRC | 09:11 | |
| *** altlogbot_1 has joined #openstack-watcher | 09:17 | |
| openstackgerrit | Dantali0n proposed openstack/watcher master: [wip] formal datasource interface implementation https://review.opendev.org/656622 | 09:29 |
| openstackgerrit | Dantali0n proposed openstack/watcher master: Allow for global datasources preference from config https://review.opendev.org/645294 | 09:40 |
| *** gmann has joined #openstack-watcher | 09:41 | |
| openstackgerrit | Dantali0n proposed openstack/watcher master: Allow for global datasources preference from config https://review.opendev.org/645294 | 09:42 |
| *** gmann is now known as gmann_pto | 09:57 | |
| *** altlogbot_1 has quit IRC | 11:19 | |
| *** adiantum has joined #openstack-watcher | 11:20 | |
| *** altlogbot_0 has joined #openstack-watcher | 11:21 | |
| openstackgerrit | sumitjami proposed openstack/watcher master: WIP allow using file to override metric map https://review.opendev.org/657374 | 11:29 |
| openstackgerrit | licanwei proposed openstack/python-watcherclient master: [WiP]Add force option https://review.opendev.org/657375 | 11:37 |
| *** altlogbot_0 has quit IRC | 12:15 | |
| *** altlogbot_2 has joined #openstack-watcher | 12:21 | |
| *** adiantum has quit IRC | 12:53 | |
| *** adiantum has joined #openstack-watcher | 13:05 | |
| *** altlogbot_2 has quit IRC | 13:15 | |
| *** altlogbot_0 has joined #openstack-watcher | 13:17 | |
| *** altlogbot_0 has quit IRC | 13:21 | |
| *** altlogbot_1 has joined #openstack-watcher | 13:27 | |
| *** ianychoi has quit IRC | 13:48 | |
| *** ianychoi has joined #openstack-watcher | 13:52 | |
| *** adisky__ has quit IRC | 13:55 | |
| *** ianychoi has quit IRC | 14:04 | |
| *** adiantum has quit IRC | 14:11 | |
| *** adiantum has joined #openstack-watcher | 14:12 | |
| *** mriedem has joined #openstack-watcher | 14:14 | |
| mriedem | Dantalion: hi, i attended jose and spyros' talk last week about how cern is using watcher for resource utilization https://www.openstack.org/summit/denver-2019/summit-schedule/events/23187/improving-resource-availability-in-cern-private-cloud and had some questions if you have a minute | 14:15 |
| mriedem | we (huawei) are having the same issues with scaling watcher (multi-cell nova deployment, thousands of hosts in the public cloud) | 14:16 |
| mriedem | i was looking at the cern watcher grafana feature branch https://gitlab.cern.ch/cloud-infrastructure/watcher/commits/grafana and it looks like some of that is translating to upstream work, | 14:17 |
| mriedem | 1. https://specs.openstack.org/openstack/watcher-specs/specs/train/approved/formal-datasource-interface.html | 14:17 |
| mriedem | 2. https://specs.openstack.org/openstack/watcher-specs/specs/train/approved/global-datasource-preference.html | 14:17 |
| mriedem | https://review.opendev.org/#/c/649341/ | 14:17 |
| openstackgerrit | sumitjami proposed openstack/watcher master: WIP allow using file to override metric map https://review.opendev.org/657374 | 14:18 |
| *** ianychoi has joined #openstack-watcher | 14:18 | |
| mriedem | i was wondering if there was anything upstream yet about restricting decisions to specific nova cells yet to avoid needing to look at an entire nova deployment to make a scheduling decision (since you currently can't migrate a server across cells in nova) | 14:18 |
| mriedem | althogh i'm working on making that happen for resize/cold-migrate https://specs.openstack.org/openstack/nova-specs/specs/train/approved/cross-cell-resize.html | 14:19 |
| Dantalion | mriedem: The current methodoly to prevent calling nova.hypervisors.list() is by supplying an audit scope these can contain aggregates and availability zones | 15:04 |
| Dantalion | mriedem: This spec explains how an audit scope limits the amount of external service calls http://specs.openstack.org/openstack/watcher-specs/specs/stein/implemented/scope-for-watcher-datamodel.html | 15:09 |
| mriedem | ah great, was just looking at the train specs | 15:09 |
| mriedem | Dantalion: also i just reviewed https://review.opendev.org/#/c/645294/ - hopefully i'm not overstepping on that one :) | 15:09 |
| Dantalion | mriedem: as long as your cells are seperated by either availability zones or aggregates you can use it to prevent crossing cell boundaries :) | 15:10 |
| mriedem | yup makes sense, and aggregates is how you'd group all hosts in a cell in nova | 15:10 |
| mriedem | with cells v2 anyway | 15:10 |
| mriedem | Dantalion: is there a spec or anything written up somewhere on the grafana front end you're doing? | 15:11 |
| Dantalion | mriedem: No not yet as its implementation requires on other specs to be implemented to be done properly | 15:11 |
| mriedem | my understanding is cern uses collectd to pump data to a series of influxdb nodes which is then fronted by grafana for watcher | 15:11 |
| mriedem | ok i'm guessing those are your 2 specs for train | 15:12 |
| Dantalion | mriedem: that is correct collectd -> influxdb -> grafana, this method gives you a single endpoint (grafana) to configure the datasources | 15:12 |
| Dantalion | mriedem: yes the two specs for train and also the file based metric map for which the spec was recently approved will have an impact on how I need to implement the grafana datasource | 15:13 |
| mriedem | yup, cool | 15:13 |
| mriedem | ok i'll need to read https://specs.openstack.org/openstack/watcher-specs/specs/train/approved/file-based-metricmap.html | 15:14 |
| mriedem | i'm pretty much completely new to watcher so this is all new to me | 15:14 |
| mriedem | thanks for being patient | 15:14 |
| Dantalion | not a problem I am here in this channel during office hours (+2 UTC) for questions and further discussions | 15:15 |
| *** adiantum has quit IRC | 15:27 | |
| *** adiantum has joined #openstack-watcher | 15:27 | |
| *** adiantum has quit IRC | 15:27 | |
| *** gmann_pto has quit IRC | 16:56 | |
| *** adiantum has joined #openstack-watcher | 21:19 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!