| *** zhuangl_ has joined #openstack-telemetry | 00:10 | |
| *** yuntongjin has joined #openstack-telemetry | 00:26 | |
| *** yuntongjin has left #openstack-telemetry | 00:27 | |
| *** yuntongjin1 has joined #openstack-telemetry | 00:37 | |
| *** chlong has quit IRC | 00:51 | |
| *** david-lyle has quit IRC | 00:54 | |
| *** david-lyle has joined #openstack-telemetry | 00:54 | |
| *** dave-mccowan has quit IRC | 01:01 | |
| *** cheneydc has joined #openstack-telemetry | 01:02 | |
| openstackgerrit | liusheng proposed openstack/python-aodhclient: Don't support "alarm search" and mandatory --type in "alarm list" https://review.openstack.org/283958 | 01:03 |
|---|---|---|
| *** chlong has joined #openstack-telemetry | 01:05 | |
| *** samstudi18 is now known as samstudio8 | 01:13 | |
| openstackgerrit | liusheng proposed openstack/python-aodhclient: Fix the -q/--query in threshold alarm creation https://review.openstack.org/289991 | 01:18 |
| *** _nadya_ has joined #openstack-telemetry | 01:45 | |
| *** _nadya_ has quit IRC | 01:50 | |
| *** thorst has joined #openstack-telemetry | 02:00 | |
| *** thorst has quit IRC | 02:05 | |
| *** chlong has quit IRC | 02:07 | |
| *** caisan has joined #openstack-telemetry | 02:19 | |
| *** chlong has joined #openstack-telemetry | 02:20 | |
| *** zhuangl__ has joined #openstack-telemetry | 02:30 | |
| *** zhuangl_ has quit IRC | 02:33 | |
| *** ljxiash_ has joined #openstack-telemetry | 02:50 | |
| *** ljxiash_ has quit IRC | 02:51 | |
| *** ljxiash_ has joined #openstack-telemetry | 02:56 | |
| *** dave-mccowan has joined #openstack-telemetry | 03:06 | |
| *** ljxiash_ has quit IRC | 03:07 | |
| *** ljxiash_ has joined #openstack-telemetry | 03:07 | |
| *** chlong has quit IRC | 03:22 | |
| *** dave-mccowan has quit IRC | 03:26 | |
| *** ljxiash_ has quit IRC | 03:27 | |
| *** chlong has joined #openstack-telemetry | 03:41 | |
| *** caisan has quit IRC | 03:42 | |
| *** caisan has joined #openstack-telemetry | 03:43 | |
| *** _nadya_ has joined #openstack-telemetry | 03:46 | |
| *** links has joined #openstack-telemetry | 03:50 | |
| *** _nadya_ has quit IRC | 03:50 | |
| *** yuntongjin1 has quit IRC | 03:52 | |
| *** ekarlso- has quit IRC | 04:00 | |
| *** ekarlso- has joined #openstack-telemetry | 04:13 | |
| *** thorst has joined #openstack-telemetry | 04:27 | |
| *** thorst has quit IRC | 04:35 | |
| *** chlong has quit IRC | 04:35 | |
| *** chlong has joined #openstack-telemetry | 04:47 | |
| *** yuntongjin has joined #openstack-telemetry | 04:56 | |
| *** _nadya_ has joined #openstack-telemetry | 05:02 | |
| *** chlong has quit IRC | 05:03 | |
| *** boris-42 has joined #openstack-telemetry | 05:05 | |
| *** _nadya_ has quit IRC | 05:07 | |
| *** chlong has joined #openstack-telemetry | 05:16 | |
| *** vishwana_ has joined #openstack-telemetry | 05:16 | |
| *** vishwanathj_away has quit IRC | 05:18 | |
| *** chlong has quit IRC | 05:22 | |
| *** thorst has joined #openstack-telemetry | 05:33 | |
| *** chlong has joined #openstack-telemetry | 05:34 | |
| *** thorst has quit IRC | 05:41 | |
| *** yprokule has joined #openstack-telemetry | 05:41 | |
| *** ev1lkry has quit IRC | 05:44 | |
| *** vishwana_ is now known as vishwanathj_away | 05:58 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/python-aodhclient: Add --query to alarm list https://review.openstack.org/291541 | 06:09 |
| *** pcaruana has quit IRC | 06:39 | |
| *** liamji has quit IRC | 06:47 | |
| *** ljxiash has quit IRC | 06:48 | |
| *** chlong has quit IRC | 06:48 | |
| *** thorst has joined #openstack-telemetry | 07:02 | |
| *** guillaume_ has joined #openstack-telemetry | 07:05 | |
| *** thorst has quit IRC | 07:07 | |
| *** yuntongjin1 has joined #openstack-telemetry | 07:10 | |
| *** hughhalf_ has joined #openstack-telemetry | 07:14 | |
| *** yuntongjin has quit IRC | 07:15 | |
| *** zhuangl__ has quit IRC | 07:15 | |
| *** sileht has quit IRC | 07:15 | |
| *** hughhalf has quit IRC | 07:15 | |
| *** elemoine has quit IRC | 07:15 | |
| *** g3ek has quit IRC | 07:15 | |
| *** guillaume_ is now known as g3ek | 07:15 | |
| *** zhuangl__ has joined #openstack-telemetry | 07:16 | |
| *** elemoine has joined #openstack-telemetry | 07:16 | |
| *** sileht has joined #openstack-telemetry | 07:16 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Add some resource types tests https://review.openstack.org/270419 | 07:22 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute number https://review.openstack.org/270091 | 07:22 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute bool https://review.openstack.org/270418 | 07:22 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute uuid https://review.openstack.org/270090 | 07:22 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Don't create Ceilometer resource types by default. https://review.openstack.org/270322 | 07:22 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Move legacy Ceilometer resource into indexer. https://review.openstack.org/270266 | 07:22 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource type CRUD. https://review.openstack.org/269844 | 07:22 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute string https://review.openstack.org/269888 | 07:22 |
| *** yuntongjin1 has quit IRC | 07:24 | |
| *** yuntongjin has joined #openstack-telemetry | 07:24 | |
| openstackgerrit | liusheng proposed openstack/python-aodhclient: Add --filter to "alarm list" https://review.openstack.org/292233 | 07:26 |
| *** tesseract has joined #openstack-telemetry | 07:52 | |
| *** tesseract is now known as Guest59187 | 07:52 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: devstack: allow uwsgi deployments https://review.openstack.org/292077 | 07:55 |
| *** belmoreira has joined #openstack-telemetry | 07:59 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/aodh: devstack: allow uwsgi deployments https://review.openstack.org/292245 | 07:59 |
| *** ildikov has quit IRC | 07:59 | |
| *** _nadya_ has joined #openstack-telemetry | 08:01 | |
| *** thorst has joined #openstack-telemetry | 08:07 | |
| *** thorst has quit IRC | 08:11 | |
| *** liusheng has quit IRC | 08:15 | |
| *** liusheng has joined #openstack-telemetry | 08:16 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/python-aodhclient: Trival fix a typo https://review.openstack.org/292251 | 08:24 |
| *** zhuangl__ has quit IRC | 08:25 | |
| *** rcernin has joined #openstack-telemetry | 08:25 | |
| openstackgerrit | Xia Linjuan proposed openstack/python-aodhclient: Fix aodh client fails when command with the arg --time-constraint https://review.openstack.org/272946 | 08:32 |
| *** zhuangl_ has joined #openstack-telemetry | 08:35 | |
| *** ljxiash has joined #openstack-telemetry | 08:42 | |
| *** pcaruana has joined #openstack-telemetry | 08:43 | |
| openstackgerrit | Merged openstack/python-aodhclient: Trival fix a typo https://review.openstack.org/292251 | 08:49 |
| *** permalac has joined #openstack-telemetry | 08:50 | |
| openstackgerrit | liusheng proposed openstack/python-aodhclient: Add --filter to "alarm list" https://review.openstack.org/292233 | 08:57 |
| *** ljxiash has quit IRC | 08:59 | |
| *** ljxiash has joined #openstack-telemetry | 09:00 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/aodh: devstack: allow uwsgi deployments https://review.openstack.org/292245 | 09:00 |
| *** safchain has joined #openstack-telemetry | 09:03 | |
| *** thorst has joined #openstack-telemetry | 09:09 | |
| *** zhuangl_ has quit IRC | 09:12 | |
| *** thorst has quit IRC | 09:17 | |
| *** Kevin_Zheng has joined #openstack-telemetry | 09:22 | |
| *** yuntongjin has quit IRC | 09:23 | |
| *** yassine__ has joined #openstack-telemetry | 09:27 | |
| *** yprokule has quit IRC | 09:29 | |
| *** yprokule has joined #openstack-telemetry | 09:34 | |
| *** caisan has quit IRC | 09:41 | |
| *** caisan has joined #openstack-telemetry | 09:42 | |
| *** yprokule1 has joined #openstack-telemetry | 09:46 | |
| *** yprokule has quit IRC | 09:48 | |
| *** caisan has quit IRC | 09:52 | |
| *** caisan has joined #openstack-telemetry | 09:55 | |
| *** idonotknow has joined #openstack-telemetry | 09:55 | |
| idonotknow | Hello,everyone.Does anyone know why the meter lb.active.connections is always 0...And is there any possible way to get the total number of vms in a pool? | 09:57 |
| *** cheneydc has quit IRC | 10:02 | |
| *** ljxiash has quit IRC | 10:13 | |
| *** caisan has quit IRC | 10:16 | |
| *** caisan has joined #openstack-telemetry | 10:23 | |
| *** cdent has joined #openstack-telemetry | 10:28 | |
| *** idonotknow has quit IRC | 10:36 | |
| *** thorst has joined #openstack-telemetry | 10:38 | |
| *** thorst has quit IRC | 10:43 | |
| openstackgerrit | Merged openstack/ceilometer: Fixed corner cases of incorrect use of oslo.config https://review.openstack.org/287636 | 10:43 |
| *** dave-mccowan has joined #openstack-telemetry | 10:45 | |
| *** idonotknow has joined #openstack-telemetry | 10:45 | |
| openstackgerrit | Merged openstack/python-ceilometerclient: remove default value of repeat-actions https://review.openstack.org/289238 | 10:50 |
| *** yprokule1 has quit IRC | 10:52 | |
| *** yprokule has joined #openstack-telemetry | 10:53 | |
| *** yassine__ has quit IRC | 10:55 | |
| *** mdih has joined #openstack-telemetry | 11:03 | |
| openstackgerrit | Merged openstack/python-aodhclient: Fix aodh client fails when command with the arg --time-constraint https://review.openstack.org/272946 | 11:04 |
| *** cdent has quit IRC | 11:17 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/python-aodhclient: [WIP] add name support for show and delete https://review.openstack.org/292305 | 11:20 |
| *** cdent has joined #openstack-telemetry | 11:27 | |
| *** gordc has joined #openstack-telemetry | 11:35 | |
| *** cdent_ has joined #openstack-telemetry | 11:37 | |
| *** cdent has quit IRC | 11:37 | |
| *** cdent_ is now known as cdent | 11:37 | |
| *** cdent has quit IRC | 11:41 | |
| *** cdent has joined #openstack-telemetry | 11:42 | |
| *** thorst has joined #openstack-telemetry | 11:42 | |
| *** thorst_ has joined #openstack-telemetry | 11:43 | |
| *** cdent has quit IRC | 11:45 | |
| *** thorst has quit IRC | 11:47 | |
| *** zhuangl_ has joined #openstack-telemetry | 11:51 | |
| *** cdent has joined #openstack-telemetry | 11:51 | |
| *** yassine__ has joined #openstack-telemetry | 11:57 | |
| *** chlong has joined #openstack-telemetry | 12:01 | |
| *** caisan has quit IRC | 12:03 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/python-aodhclient: Add name support for show and delete https://review.openstack.org/292305 | 12:28 |
| openstackgerrit | Zhenyu Zheng proposed openstack/python-aodhclient: Add --query to alarm list https://review.openstack.org/291541 | 12:28 |
| openstackgerrit | Zhenyu Zheng proposed openstack/python-aodhclient: Don't support "alarm search" and mandatory --type in "alarm list" https://review.openstack.org/283958 | 12:28 |
| openstackgerrit | Zhenyu Zheng proposed openstack/python-aodhclient: Add --query to alarm list https://review.openstack.org/291541 | 12:30 |
| Kevin_Zheng | ...very strange with git | 12:32 |
| *** boris-42 has quit IRC | 12:34 | |
| *** MVenesio has joined #openstack-telemetry | 12:40 | |
| openstackgerrit | Merged openstack/python-aodhclient: Fix the -q/--query in threshold alarm creation https://review.openstack.org/289991 | 12:40 |
| sileht | / | 12:41 |
| *** links has quit IRC | 12:46 | |
| *** links has joined #openstack-telemetry | 12:46 | |
| *** cdent has quit IRC | 13:01 | |
| *** cdent has joined #openstack-telemetry | 13:02 | |
| *** cdent has quit IRC | 13:06 | |
| *** zhuangl_ has quit IRC | 13:08 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: devstack: allow uwsgi deployments https://review.openstack.org/292077 | 13:16 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/aodh: devstack: allow uwsgi deployments https://review.openstack.org/292245 | 13:17 |
| *** yarkot_ has joined #openstack-telemetry | 13:19 | |
| *** peristeri has joined #openstack-telemetry | 13:23 | |
| *** BigWillie has joined #openstack-telemetry | 13:24 | |
| *** yarkot_ has quit IRC | 13:27 | |
| *** jwcroppe has joined #openstack-telemetry | 13:29 | |
| *** krotscheck has joined #openstack-telemetry | 13:32 | |
| *** cdent has joined #openstack-telemetry | 13:35 | |
| *** ljxiash has joined #openstack-telemetry | 13:41 | |
| openstackgerrit | Merged openstack/ceilometer: Remove en_GB translations https://review.openstack.org/292085 | 13:43 |
| *** rbak has quit IRC | 13:51 | |
| *** yprokule has quit IRC | 13:58 | |
| *** nicodemus_ has joined #openstack-telemetry | 13:58 | |
| nicodemus_ | morning gordc | 13:59 |
| gordc | nicodemus_: hey, whatsup? | 13:59 |
| nicodemus_ | do you have a moment? I'd like to ask you about ceilometer | 14:00 |
| gordc | sure thing | 14:00 |
| nicodemus_ | on kilo, can I limit what measures taken by the agents using the pipeline.yaml? | 14:01 |
| nicodemus_ | because I'm having a lot of messages queued on the metering.sample queue | 14:01 |
| nicodemus_ | and two collectors can't seem to handle the amount of messages | 14:02 |
| gordc | nicodemus_: when you use gnocchi backend? | 14:02 |
| *** ig0r_ has joined #openstack-telemetry | 14:02 | |
| *** yprokule has joined #openstack-telemetry | 14:03 | |
| gordc | you can/should use pipeline.yaml to dictate what meters/events are built. | 14:03 |
| nicodemus_ | gordc, yes, with gnocchi... The gnocchi API doesn't seem to be the bottleneck, plus I have more than one | 14:03 |
| nicodemus_ | with haproxy | 14:04 |
| gordc | do you have a backlog on notifications.info and notifications.sample queues as well? | 14:04 |
| gordc | if it's only the metering.sample queue. i think it's an issue with how your gnocchi api or dispatcher is configured. | 14:05 |
| *** boris-42 has joined #openstack-telemetry | 14:05 | |
| nicodemus_ | I don't have notifications.sample queue, and the notifications.info has no consumers | 14:05 |
| nicodemus_ | I'm only running ceilometer-collector and agent-compute | 14:06 |
| gordc | ... oh right. you have kilo | 14:07 |
| nicodemus_ | the dispatcher_gnocchi section only has: http://paste.openstack.org/show/490390/ | 14:07 |
| gordc | do you have caching in gnocchi dispatcher? | 14:07 |
| gordc | i would say that's a requirement | 14:08 |
| nicodemus_ | I believe not... how could I check that? | 14:08 |
| gordc | we added it in Mitaka so if you're using Liberty dispatcher, it won't be there. | 14:08 |
| gordc | nicodemus_: you shoudl have a [cache] section in your conf file. | 14:09 |
| *** tongli has joined #openstack-telemetry | 14:10 | |
| nicodemus_ | hmm... could I try using the mitaka dispatcher? | 14:10 |
| *** KrishR has joined #openstack-telemetry | 14:10 | |
| nicodemus_ | What I did was to copy both gnocchi.py and gnocchi_client.py from liberty to kilo and update the setup.cfg to include gnocchi as dispatcher | 14:11 |
| gordc | with kilo? possibly? it will bring in gnocchiclient requirement but i believe you already ahve that. | 14:11 |
| *** rbak has joined #openstack-telemetry | 14:11 | |
| gordc | ok. gnocchi_client.py isn't in Mitaka since basically that was just a placeholder until we got client complete. | 14:12 |
| tongli | @gordc, good morning, any way at all to bring in the http publisher in for this release? Since the removal of the cadf_only flag, it is kind of hard to send only cadf messages in this release. | 14:12 |
| gordc | tongli: we are pretty far past feature freeze and FFE. | 14:14 |
| gordc | i don't see how your patch resolves cadf messages? | 14:15 |
| tongli | @gordc, I know we are at the code freeze, I am aware of that. but this is really just like a plugin. | 14:15 |
| tongli | by default it is not configured. | 14:15 |
| *** david-lyle has quit IRC | 14:16 | |
| tongli | @gordc, this one alone does not, I am trying also to put in a transformer which extract piece of information | 14:16 |
| gordc | sorry. this is not a bug, which is what this period is for... FFE were suppose to requested weeks ago. | 14:16 |
| tongli | @gordc, sigh! | 14:17 |
| gordc | if it's a transformer AND a publisher, that seems like a feature | 14:17 |
| tongli | yes, it will be. | 14:17 |
| gordc | i'd also say, it's already possible to do as is. | 14:17 |
| tongli | ? I like to hear how. | 14:18 |
| gordc | you just need to push audit.* events to one collector and have that collector send http to whatever. | 14:18 |
| gordc | i'm assuming you want audit.* events | 14:18 |
| tongli | true, thought about that as well, but the message itself is not cadf message. | 14:19 |
| tongli | need extract piece of the message. | 14:19 |
| gordc | what's the message right now? | 14:19 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Add some resource types tests https://review.openstack.org/270419 | 14:19 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute number https://review.openstack.org/270091 | 14:19 |
| tongli | without this http publisher, collector is absolutely necessary. | 14:19 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute bool https://review.openstack.org/270418 | 14:19 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute uuid https://review.openstack.org/270090 | 14:19 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Don't create Ceilometer resource types by default. https://review.openstack.org/270322 | 14:19 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Move legacy Ceilometer resource into indexer. https://review.openstack.org/270266 | 14:19 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource type CRUD. https://review.openstack.org/269844 | 14:19 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute string https://review.openstack.org/269888 | 14:19 |
| *** david-lyle has joined #openstack-telemetry | 14:19 | |
| tongli | @gordc, the message buried in the cadf in resource_metadata I think. | 14:20 |
| gordc | that was 2 cycles ago | 14:20 |
| gordc | that version has been deprecated and removed already | 14:20 |
| gordc | https://github.com/openstack/keystonemiddleware/blob/master/keystonemiddleware/audit.py#L397-L401 | 14:20 |
| gordc | that's the current liberty/mitaka version. it's the reason we dropped the cadf_only stuff.. because it's not relevant anymore. | 14:21 |
| tongli | @gordc, let me verify that , I think I tried in liberty, the message still needs to be extracted to make a true cadf message. | 14:22 |
| tongli | with http publisher, we do not have to have collector running, | 14:23 |
| tongli | without it, notification agent has to place the message back to the queue, collector has to get it then post, too many trips. | 14:23 |
| gordc | tongli: if it's a problem with liberty, why are you patching something that doesn't exist in mitaka? | 14:24 |
| gordc | i'm pretty sure the cadf_only stuff exists in liberty regardless... it's just deprecated. | 14:24 |
| tongli | think mitaka worked same way. | 14:24 |
| tongli | cadf_only is in liberty, but not in mitaka | 14:24 |
| gordc | right | 14:24 |
| tongli | let me verify that cadf message does not need to be extracted in both releases. | 14:26 |
| gordc | i don't mind having http publisher. i do think the original reasoning that the synchronously HTTP req can slow down notification agent is still valid but meh | 14:26 |
| gordc | it's just way past feature freeze | 14:26 |
| tongli | @gordc, understood, | 14:26 |
| gordc | you should stop using deprecated middleware in liberty if you still see it embedded in metadata | 14:27 |
| tongli | I think I was using keystonemiddleware, that is recommended way, right? | 14:27 |
| gordc | yep | 14:27 |
| *** pcaruana has quit IRC | 14:28 | |
| tongli | let me verify and get back to you here. this is kind important to us. | 14:28 |
| gordc | sure | 14:30 |
| *** idonotknow has quit IRC | 14:36 | |
| cdent | gordc: gabbi uses httplib2 | 14:47 |
| gordc | cdent: oh? that's all blackbox to me. | 14:50 |
| cdent | good that's how it should be | 14:50 |
| cdent | test_bin is a horrible thing | 14:51 |
| cdent | you should just drop that | 14:51 |
| cdent | I'll write on the bug | 14:51 |
| MVenesio | gordc: hi Gordon, quick question, do u remember how is the way to configure ceilometer to read from a different virtual Host with user and password ? | 14:52 |
| gordc | cdent: i don't know what it does to be honest. | 14:52 |
| cdent | starts the binaries, and makes sure they are running | 14:52 |
| cdent | which is...strange | 14:52 |
| gordc | *shrugs* i'll 'accidently' delete it as part of another patch | 14:52 |
| gordc | MVenesio: yes, you can configure multiple vhost. i believe there is an example in config | 14:53 |
| MVenesio | gordc: do u remember how the flag lookslike ? | 14:56 |
| gordc | whoops sorry, got caught up in another ping... will continue diggin | 14:57 |
| gordc | https://github.com/openstack/ceilometer/blob/e397e9a519480fe6776f3cb45cd064d9f9093a74/ceilometer/notification.py#L63-L69 | 14:57 |
| MVenesio | gordc: thxxx | 14:58 |
| gordc | this is sileht's example https://review.openstack.org/gitweb?p=openstack%2Fceilometer.git;a=commitdiff;h=3d3480a813c7da788beb6440b719f3dfe895ef3a | 14:59 |
| MVenesio | gordc: so, seems like messaging_urls its not prepared to support rabbit clusters | 15:03 |
| MVenesio | gordc: sorry, yes it is | 15:04 |
| MVenesio | gordc: i'll test it, thanks | 15:05 |
| *** yprokule has quit IRC | 15:10 | |
| stevelle | Has there been any talk about Newton cycle priorities in the telemetry project space? I haven't seen a Newton-specific etherpad pop up or the topic in the meetings yet. | 15:11 |
| *** yprokule has joined #openstack-telemetry | 15:15 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Add some resource types tests https://review.openstack.org/270419 | 15:16 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute number https://review.openstack.org/270091 | 15:16 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute bool https://review.openstack.org/270418 | 15:16 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute uuid https://review.openstack.org/270090 | 15:16 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Don't create Ceilometer resource types by default. https://review.openstack.org/270322 | 15:16 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Move legacy Ceilometer resource into indexer. https://review.openstack.org/270266 | 15:16 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource type CRUD. https://review.openstack.org/269844 | 15:16 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: Implements resource attribute string https://review.openstack.org/269888 | 15:16 |
| *** links has quit IRC | 15:23 | |
| *** ig0r__ has joined #openstack-telemetry | 15:23 | |
| *** ig0r_ has quit IRC | 15:26 | |
| *** yprokule has quit IRC | 15:28 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: gate: add missing sudo https://review.openstack.org/292456 | 15:28 |
| openstackgerrit | gordon chung proposed openstack/ceilometer: improve notification processing https://review.openstack.org/275741 | 15:39 |
| *** liamji has joined #openstack-telemetry | 15:46 | |
| *** thorst_ is now known as thorst_afk | 15:46 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: workaround to _strptime import issue on py2 https://review.openstack.org/292465 | 15:48 |
| openstackgerrit | Julien Danjou proposed openstack/ceilometer: tests: remove ceilometer-api bin test cases https://review.openstack.org/292475 | 16:00 |
| stevelle | jd__: re my above query for Newton, I believe there was something about Rally on the roadmap for Gnocci on the roadmap. Is there any other details on that? | 16:05 |
| jd__ | stevelle: no idea, I never heard of that – though I don't follow Rally | 16:05 |
| boris-42 | jd__: you don't follow rally how soooo?) | 16:06 |
| jd__ | :-) | 16:06 |
| stevelle | ah, just checked again. my mistake it was on ceilometer https://wiki.openstack.org/wiki/Telemetry/RoadMap | 16:06 |
| stevelle | sry jd__ | 16:06 |
| boris-42 | stevelle: btw I heard some guys were developing rally benchmarks for gnocci | 16:07 |
| boris-42 | stevelle: unfortunately I don't remember nick names=( it was like 3 months ago | 16:08 |
| stevelle | boris-42: yes, I'm helping someone work on rally benchmarks for ceilometer + gnocchi together, which is primarily about benchmarking ceilometer right now. You talked to me about it. | 16:08 |
| boris-42 | stevelle: ah that was you great=) | 16:09 |
| boris-42 | stevelle: if you need any future help from rally side ping me | 16:09 |
| stevelle | smurke is doing the primary work on that | 16:09 |
| jd__ | would be cool to have Gnocchi benchmarks without Ceilometer though :) | 16:09 |
| *** pcaruana has joined #openstack-telemetry | 16:09 | |
| stevelle | jd__: I have an Ansible role to deploy Gnocchi in progress and it should merge to openstack-ansible early in Newton | 16:10 |
| stevelle | for benchmarking you are primarily interested in comparing storage engines jd__? | 16:11 |
| *** cdent has quit IRC | 16:12 | |
| *** _nadya_ has quit IRC | 16:31 | |
| *** Guest59187 has quit IRC | 16:31 | |
| *** cdent has joined #openstack-telemetry | 16:39 | |
| *** ljxiash has quit IRC | 16:41 | |
| *** liamji has quit IRC | 16:43 | |
| *** rcernin has quit IRC | 16:56 | |
| *** chlong has quit IRC | 16:56 | |
| nicodemus_ | Is it correct to assume that each "measure_" object in the backend is created by each POST done to the gnocchi api? (a POST to /v1/resource/.../measures) | 17:00 |
| nicodemus_ | I'm trying to profile the amount of objects that are created in the ceph pool, and the amount of such measures that are processed in the same timeframe | 17:00 |
| *** rbak_ has joined #openstack-telemetry | 17:01 | |
| *** rbak has quit IRC | 17:05 | |
| *** mdih has quit IRC | 17:06 | |
| openstackgerrit | Julien Danjou proposed openstack/python-aodhclient: Use overtest to run Gnocchi in tests https://review.openstack.org/292515 | 17:06 |
| *** belmoreira has quit IRC | 17:07 | |
| openstackgerrit | Julien Danjou proposed openstack/python-aodhclient: tox: remove useless install_command https://review.openstack.org/292517 | 17:07 |
| *** yassine__ has quit IRC | 17:15 | |
| *** thorst_afk is now known as thorst_ | 17:20 | |
| *** ildikov_ has quit IRC | 17:25 | |
| gordc | nicodemus_: yes, each measure object roughly represents each POST api request (unless you use batch write interface) | 17:30 |
| nicodemus_ | gordc, cool. Thanks! | 17:31 |
| MVenesio | sileht: i'm testing the messaging_urls flag to consume messages from different vhosts i tested this rabbit://cinder:CuRcg17Mry1tj8ATh7u@10.10.10.35:5672/cinder but i'm getting an error with invalid credentials, user, url and pass are tested and working for cinder, there's something wrong ? | 17:39 |
| *** _nadya_ has joined #openstack-telemetry | 17:41 | |
| openstackgerrit | Merged openstack/ceilometer: gate: add missing sudo https://review.openstack.org/292456 | 17:49 |
| *** _nadya_ has quit IRC | 17:49 | |
| *** jmlowe1 has quit IRC | 17:50 | |
| *** safchain has quit IRC | 17:51 | |
| *** alejandrito has joined #openstack-telemetry | 17:51 | |
| *** flwang1 has joined #openstack-telemetry | 17:59 | |
| *** _nadya_ has joined #openstack-telemetry | 18:03 | |
| *** _nadya_ has quit IRC | 18:04 | |
| *** _nadya_ has joined #openstack-telemetry | 18:11 | |
| *** ildikov_ has joined #openstack-telemetry | 18:12 | |
| *** ildikov has joined #openstack-telemetry | 18:16 | |
| *** permalac has quit IRC | 18:19 | |
| EmilienM | sileht: thx a lot for https://review.openstack.org/290626 | 18:19 |
| *** flwang1 has quit IRC | 18:25 | |
| *** cdent has quit IRC | 18:37 | |
| *** belmoreira has joined #openstack-telemetry | 18:38 | |
| *** jwcroppe has quit IRC | 18:45 | |
| *** belmoreira has quit IRC | 18:48 | |
| nicodemus_ | gordc, do you have a moment? | 18:54 |
| *** cdent has joined #openstack-telemetry | 18:55 | |
| *** _nadya_ has quit IRC | 19:13 | |
| *** _nadya_ has joined #openstack-telemetry | 19:18 | |
| *** _nadya_ has quit IRC | 19:20 | |
| *** belmoreira has joined #openstack-telemetry | 19:36 | |
| nicodemus_ | jd__, do you have a minute? | 19:47 |
| *** flwang1 has joined #openstack-telemetry | 19:55 | |
| gordc | nicodemus_: back now. | 19:57 |
| nicodemus_ | gordc, I'm having an issue with ceilometer-collector, wanted to know if you maybe had some pointers | 20:03 |
| nicodemus_ | there's a lot of messages in the metering.sample queue, and the collectors connect to the queue and after a few minutes get disconnected | 20:04 |
| *** ig0r_ has joined #openstack-telemetry | 20:04 | |
| nicodemus_ | regardless of the amount of workers I configure in ceilometer.conf, I see them all as consumers of the queue | 20:04 |
| nicodemus_ | but then... one by one they disconnect | 20:04 |
| nicodemus_ | have you ever come across such behavior? | 20:05 |
| EmilienM | gordc, pradk: I might found a bug in Aodh Listener but I'm not sure. In our Puppet CI, it happens Aodh Listener starts before RabbitMQ user/permissions are set, but never try again later, the process crash. http://logs.openstack.org/99/286899/60/check/gate-puppet-openstack-integration-scenario001-tempest-dsvm-centos7/069651c/logs/aodh/notifier.txt.gz#_2016-03-14_17_13_14_909 | 20:05 |
| EmilienM | in my experience, services don't crash, but retry later, isn't? | 20:05 |
| gordc | nicodemus_: what oslo.messaging version are you using? do you have any logs? | 20:06 |
| nicodemus_ | oslo.messaging is 1.8.3 | 20:07 |
| *** ig0r__ has quit IRC | 20:07 | |
| gordc | EmilienM: open it. i have no idea tbh. | 20:07 |
| nicodemus_ | after the disconnection, the ceilometer-collector log keeps on writing lines about posting measures, and I see those POST on gnocchi-api | 20:07 |
| EmilienM | gordc: ok, just making sure my statement is true about timestamps | 20:08 |
| pradk | EmilienM, seems like the notifier is throwing that error , perhaps we could have a retry loop there | 20:08 |
| gordc | i didn't read it.lol | 20:08 |
| gordc | EmilienM: oh, it's the aodh-notifier? dammit. i broke it.lol | 20:09 |
| EmilienM | again? | 20:09 |
| gordc | meh | 20:09 |
| EmilienM | I hope next PTL won't break Aodh Listener | 20:10 |
| * EmilienM runs | 20:10 | |
| gordc | i don't understand irc enough to kick/ban | 20:11 |
| EmilienM | pradk: yes I confirm the timestamp | 20:11 |
| EmilienM | I'm reporting a bug in launchpad | 20:11 |
| *** KrishR has quit IRC | 20:12 | |
| *** MVenesio has quit IRC | 20:12 | |
| gordc | nicodemus_: it disconnects but still reads and writes to gnocchi? | 20:13 |
| gordc | it might just be an oslo.messaging thing and it recovers | 20:13 |
| EmilienM | pradk: https://bugs.launchpad.net/aodh/+bug/1557154 | 20:13 |
| openstack | EmilienM: Error: Could not gather data from Launchpad for bug #1557154 (https://launchpad.net/bugs/1557154). The error has been logged | 20:13 |
| nicodemus_ | apparently.. we're thinking if maybe it's caching something, because if I look at the current connections on the queue master node using netstat, I no longer see it connected | 20:15 |
| gordc | nicodemus_: and you don't have any consumers on metering.sample queue? | 20:36 |
| nicodemus_ | gordc, no | 20:37 |
| nicodemus_ | gordc, I'm trying now with a different oslo.messaging version | 20:37 |
| *** peristeri has quit IRC | 20:40 | |
| gordc | nicodemus_: i think you're already on the latest. | 20:41 |
| gordc | https://github.com/openstack/oslo.messaging/commit/43425ab3bb76ba84d4aaf3ec68c018de891f4183 | 20:41 |
| gordc | this patch looks relevant thought | 20:41 |
| gordc | nicodemus_: i'm going to request a new release for oslo.messaging kilo... there are quite a few backports that never got released. | 20:42 |
| nicodemus_ | gordc, that'd be great | 20:42 |
| *** BigWillie has quit IRC | 20:46 | |
| *** bapalm has quit IRC | 20:55 | |
| *** bapalm has joined #openstack-telemetry | 20:58 | |
| *** safchain has joined #openstack-telemetry | 21:04 | |
| *** chlong has joined #openstack-telemetry | 21:14 | |
| *** jdandrea_ is now known as jdandrea | 21:25 | |
| *** nicodemus_ has quit IRC | 21:27 | |
| *** thorst_ has quit IRC | 21:28 | |
| *** thorst_ has joined #openstack-telemetry | 21:31 | |
| openstackgerrit | Merged openstack/gnocchi: workaround to _strptime import issue on py2 https://review.openstack.org/292465 | 21:31 |
| *** thorst_ has quit IRC | 21:35 | |
| *** belmoreira has quit IRC | 21:44 | |
| *** safchain has quit IRC | 21:56 | |
| *** jdandrea_ has joined #openstack-telemetry | 21:57 | |
| *** jdandrea has quit IRC | 21:57 | |
| *** r-mibu has quit IRC | 22:20 | |
| *** r-mibu has joined #openstack-telemetry | 22:20 | |
| *** nicodemus_ has joined #openstack-telemetry | 22:24 | |
| nicodemus_ | gordc, is there a way to configure ceilometer-collector not to prefetch messages in the metering.sample queue? | 22:25 |
| nicodemus_ | each worker does a quite aggresive prefetch, and it seems to prevent the remaining workers from processing messages (unless there's thousands of messages in the queue) | 22:28 |
| gordc | nicodemus_: not quite sure what you mean by prefetch | 22:28 |
| nicodemus_ | suppose there are 200 messages waiting in the queue, when a collector worker connects to the queue it kind of "reserves" those 200 messages for processing | 22:29 |
| nicodemus_ | if a second worker connects to the queue, it has no "ready" messages to process | 22:29 |
| gordc | oh. you want it to grab less messages at a time? | 22:30 |
| gordc | so by default, oslo.messaging has 64 threads, so each worker will grab 64 for at a time. | 22:31 |
| gordc | there's a way to control that. (let me find option) | 22:31 |
| gordc | https://github.com/openstack/oslo.messaging/blob/a89321f2511a5f7f8c1de617be20718cb85d4272/oslo_messaging/server.py#L50-L55 | 22:33 |
| gordc | you can set that lower... it will probably slow down performance quite a bit though unless you use mitaka+batching | 22:34 |
| *** cdent has quit IRC | 22:36 | |
| nicodemus_ | gordc, thanks! I'll give it a try | 22:40 |
| nicodemus_ | Right now I have three ceilometer-collector services on three hw hosts, each with 10 workers... and cannot process messages fast enough | 22:41 |
| nicodemus_ | the queue is building up | 22:42 |
| *** alejandrito has quit IRC | 22:42 | |
| nicodemus_ | I figured that maybe if I have many workers each grabbing a tiny quantity of messages each, it could better parallelize processing | 22:44 |
| gordc | nicodemus_: ... you have 30 workers? um how many instances are you monitoring? | 22:47 |
| gordc | i don't think this is a collector issue. it seems to definitely be write issue with gnocchi. how many threads/processes do you have configured for gnocchi api? | 22:48 |
| nicodemus_ | 60 instances across 4 compute nodes | 22:49 |
| nicodemus_ | plus ceilometer-agent-central polling through SNMP those 4 computes | 22:49 |
| gordc | yeah, it's a gnocchi api thing. if all the messages are 'reserved' off queue, then it means the workers are waiting around trying to write to gnocchi api | 22:49 |
| nicodemus_ | I have two gnocchi-api instances behind apache wid mod-wsgi using gnocchi/rest/app.wsgi | 22:51 |
| stevelle | I'm wondering if the mod_wsgi is our best option at this point | 22:52 |
| nicodemus_ | configured for 3 processes and 32 threads (each) | 22:52 |
| nicodemus_ | I don't see the api instances struggling to cope (meaning, no high cpu/memory usage) | 22:53 |
| nicodemus_ | both do write on the same ceph pool that metricd is working on | 22:54 |
| gordc | i would bump that up: usually it's 1.5 * <#ofcores> processes | 22:56 |
| gordc | maybe change to 8 processes and/or 64 threads | 22:56 |
| gordc | right now you have 3*10*64 'parallel' threads on the collector trying to call 3*32 threads in gnocchi api | 22:57 |
| nicodemus_ | gordc, haven't thought about that... I'll give it a try as well | 22:57 |
| *** hughhalf_ is now known as hughhalf | 22:57 | |
| gordc | i'd probably turn down number of collector workers. seems high but you can play around with it. | 22:58 |
| *** rbak_ has quit IRC | 23:01 | |
| *** pradk has quit IRC | 23:02 | |
| *** gordc has quit IRC | 23:11 | |
| *** thorst_ has joined #openstack-telemetry | 23:32 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!