| *** nealph has quit IRC | 00:05 | |
| *** eglynn has quit IRC | 00:14 | |
| *** sbfox has quit IRC | 00:29 | |
| openstackgerrit | Merged openstack/ceilometer: Enable pep8 on ./tools directory https://review.openstack.org/136781 | 00:32 |
|---|---|---|
| openstackgerrit | Merged openstack/ceilometer: Fix order of arguments in assertEqual https://review.openstack.org/134940 | 00:33 |
| *** ViswaV_ has quit IRC | 00:38 | |
| *** mc__ has joined #openstack-ceilometer | 00:39 | |
| *** ViswaV has joined #openstack-ceilometer | 00:39 | |
| *** julim has joined #openstack-ceilometer | 00:40 | |
| *** ViswaV has quit IRC | 00:44 | |
| openstackgerrit | Merged openstack/ceilometer: Remove Python 2.6 classifier https://review.openstack.org/137114 | 00:49 |
| *** alexpilotti has quit IRC | 00:51 | |
| *** amalagon has joined #openstack-ceilometer | 01:06 | |
| *** Kennan has joined #openstack-ceilometer | 01:19 | |
| *** promulo__ has joined #openstack-ceilometer | 01:29 | |
| *** r-daneel has quit IRC | 01:32 | |
| *** promulo has quit IRC | 01:32 | |
| *** _cjones_ has quit IRC | 02:08 | |
| *** mc__ has quit IRC | 02:10 | |
| *** mc__2 has joined #openstack-ceilometer | 02:11 | |
| *** mc__2 is now known as mc__ | 02:11 | |
| *** promulo has joined #openstack-ceilometer | 02:14 | |
| *** nosnos has joined #openstack-ceilometer | 02:14 | |
| *** promulo__ has quit IRC | 02:17 | |
| *** sbfox has joined #openstack-ceilometer | 02:32 | |
| *** sbfox has quit IRC | 02:44 | |
| *** sbfox has joined #openstack-ceilometer | 02:59 | |
| *** nosnos has quit IRC | 03:01 | |
| *** fnaval has quit IRC | 03:07 | |
| *** sbfox has quit IRC | 03:14 | |
| *** ryanpetrello has joined #openstack-ceilometer | 03:18 | |
| *** fnaval has joined #openstack-ceilometer | 03:23 | |
| *** fnaval has quit IRC | 03:23 | |
| *** zqfan has quit IRC | 03:24 | |
| *** zqfan has joined #openstack-ceilometer | 03:25 | |
| *** harlowja is now known as harlowja_away | 03:29 | |
| *** _cjones_ has joined #openstack-ceilometer | 03:36 | |
| *** _cjones_ has quit IRC | 03:37 | |
| *** _cjones_ has joined #openstack-ceilometer | 03:38 | |
| *** ryanpetrello has quit IRC | 03:41 | |
| *** promulo__ has joined #openstack-ceilometer | 04:00 | |
| *** promulo has quit IRC | 04:02 | |
| *** sbfox has joined #openstack-ceilometer | 04:03 | |
| *** ryanpetrello has joined #openstack-ceilometer | 04:07 | |
| *** nosnos has joined #openstack-ceilometer | 04:16 | |
| *** sbfox has quit IRC | 04:18 | |
| *** _cjones_ has quit IRC | 04:22 | |
| *** ryanpetrello has quit IRC | 04:23 | |
| *** asalkeld has joined #openstack-ceilometer | 04:24 | |
| *** deepthi has joined #openstack-ceilometer | 04:30 | |
| *** rainmantea has joined #openstack-ceilometer | 04:36 | |
| *** mc__ has quit IRC | 04:37 | |
| *** yatin has joined #openstack-ceilometer | 04:42 | |
| *** sbfox has joined #openstack-ceilometer | 04:43 | |
| *** ryanpetrello has joined #openstack-ceilometer | 04:49 | |
| *** nellysmitt has joined #openstack-ceilometer | 05:03 | |
| *** nellysmitt has quit IRC | 05:08 | |
| *** _cjones_ has joined #openstack-ceilometer | 05:09 | |
| *** ishant has joined #openstack-ceilometer | 05:12 | |
| *** _cjones_ has quit IRC | 05:13 | |
| *** mc__ has joined #openstack-ceilometer | 05:21 | |
| *** sbfox has quit IRC | 05:27 | |
| *** _cjones_ has joined #openstack-ceilometer | 05:28 | |
| *** Longgeek has joined #openstack-ceilometer | 05:38 | |
| *** amalagon has quit IRC | 05:43 | |
| *** Longgeek has quit IRC | 05:44 | |
| *** sbfox has joined #openstack-ceilometer | 05:57 | |
| *** _cjones_ has quit IRC | 05:59 | |
| *** k4n0 has joined #openstack-ceilometer | 06:03 | |
| *** ryanpetrello has quit IRC | 06:05 | |
| *** amalagon has joined #openstack-ceilometer | 06:07 | |
| *** _cjones_ has joined #openstack-ceilometer | 06:08 | |
| *** Longgeek has joined #openstack-ceilometer | 06:22 | |
| *** deepthi has quit IRC | 06:35 | |
| *** deepthi has joined #openstack-ceilometer | 06:47 | |
| *** nellysmitt has joined #openstack-ceilometer | 07:04 | |
| *** nellysmitt has quit IRC | 07:09 | |
| *** exploreshaifali has joined #openstack-ceilometer | 07:18 | |
| *** sbfox has quit IRC | 07:22 | |
| *** eglynn has joined #openstack-ceilometer | 07:26 | |
| *** IvanBerezovskiy has joined #openstack-ceilometer | 07:28 | |
| rainmantea | eglynn: just wanted you to have a look at this entry at compute.log Obtaining CPU Util is not implemented for LibvirtInspector | 07:29 |
| rainmantea | eglynn: and "Hello"... | 07:30 |
| *** ala_ has joined #openstack-ceilometer | 07:37 | |
| *** ildikov has joined #openstack-ceilometer | 07:39 | |
| *** alla__ has joined #openstack-ceilometer | 07:41 | |
| *** ala_ has quit IRC | 07:44 | |
| *** eglynn has quit IRC | 07:46 | |
| *** jaypipes has quit IRC | 07:50 | |
| *** mitz- has joined #openstack-ceilometer | 07:51 | |
| *** ryanpetrello has joined #openstack-ceilometer | 07:51 | |
| *** mitz_ has quit IRC | 07:51 | |
| *** ryanpetrello has quit IRC | 07:56 | |
| *** nellysmitt has joined #openstack-ceilometer | 07:57 | |
| *** zqfan has quit IRC | 07:58 | |
| *** zqfan has joined #openstack-ceilometer | 07:59 | |
| *** ildikov has quit IRC | 08:07 | |
| *** _cjones_ has quit IRC | 08:15 | |
| *** ifarkas has joined #openstack-ceilometer | 08:16 | |
| *** eglynn has joined #openstack-ceilometer | 08:19 | |
| *** ildikov has joined #openstack-ceilometer | 08:22 | |
| rainmantea | eglynn,ildokov,nsage: I got the meters!!!yeaaaah | 08:23 |
| *** nadya has joined #openstack-ceilometer | 08:26 | |
| *** nadya is now known as Guest28381 | 08:27 | |
| *** ifarkas has quit IRC | 08:28 | |
| ildikov | rainmantea: \o/ :) | 08:28 |
| ildikov | rainmantea: what was the solution finally? | 08:28 |
| rainmantea | guess what---- # AMQP exchange to connect to if using RabbitMQ or Qpid # (string value) control_exchange=openstack | 08:30 |
| rainmantea | this had to be uncommented | 08:30 |
| rainmantea | now i know its preliminary troubleshoot.. but all i say is It wasnt in the DOC!! | 08:31 |
| rainmantea | in the ceilometer.conf file | 08:31 |
| rainmantea | ildikov:i know u must be thinking "whaat u didnt do that before..." //sigh | 08:32 |
| rainmantea | annnnyway im off to auto scaling n stuff finally (damn couldnt get alarms to conjure up coz of missing meters..lol) | 08:32 |
| ildikov | rainmantea: you mean that you didn't change the default value, just uncommented that line? | 08:32 |
| rainmantea | yes | 08:33 |
| openstackgerrit | Lan Qi Song proposed openstack/ceilometer: Database.max_retries only override on sqlalchemy side https://review.openstack.org/136964 | 08:33 |
| *** mitz_ has joined #openstack-ceilometer | 08:34 | |
| ildikov | rainmantea: hmm, interesting, what I saw earlier in the ceilometer.conf to be set is the notifictaion_topics, I didn't remember that the control_exchange has to be explicitly set... | 08:35 |
| rainmantea | ildikov:yeah | 08:36 |
| *** mitz- has quit IRC | 08:36 | |
| rainmantea | ildikov: though my ceilometer.conf holds these also uncommented as i was doing all possible combinations | 08:37 |
| rainmantea | # Exchanges name to listen for notifications. (multi valued) http_control_exchanges=nova http_control_exchanges=glance http_control_exchanges=neutron http_control_exchanges=cinder | 08:37 |
| ildikov | rainmantea: a-ha, I see, well, I will check the code later, if I will have time today to see how the default values are configured | 08:40 |
| ildikov | rainmantea: have you changed anything else | 08:40 |
| ildikov | rainmantea: ? | 08:40 |
| rainmantea | iildikov: well i commented the hypervisor : libvirt and domain: qemu option | 08:41 |
| rainmantea | also | 08:41 |
| *** zqfan has quit IRC | 08:41 | |
| rainmantea | they were commented before..i uncommented to see results some time ag(2 days i think ago..) | 08:42 |
| *** zqfan has joined #openstack-ceilometer | 08:42 | |
| *** safchain has joined #openstack-ceilometer | 08:54 | |
| openstackgerrit | Igor Degtiarov proposed openstack/ceilometer: [MongoDB] Fix bug with 'bad' chars in metadatas keys https://review.openstack.org/121003 | 08:56 |
| openstackgerrit | Igor Degtiarov proposed openstack/ceilometer: Clean unused tables from mongodb and db2 https://review.openstack.org/132561 | 08:58 |
| *** promulo__ has quit IRC | 09:14 | |
| ildikov | rainmantea: a-ha, ok | 09:15 |
| ildikov | rainmantea: didn't you say something about you're running your deployment above VMware? | 09:16 |
| openstackgerrit | mizeng proposed openstack/ceilometer: fix for https://bugs.launchpad.net/ceilometer/+bug/1396473 https://review.openstack.org/137308 | 09:20 |
| openstackgerrit | ZhiQiang Fan proposed openstack/python-ceilometerclient: Add apiclient to openstack-common.conf https://review.openstack.org/137309 | 09:24 |
| *** mc__ has quit IRC | 09:31 | |
| *** cmyster has quit IRC | 09:35 | |
| *** cmyster has joined #openstack-ceilometer | 09:36 | |
| *** Longgeek_ has joined #openstack-ceilometer | 09:38 | |
| *** Longgeek has quit IRC | 09:39 | |
| *** zqfan has quit IRC | 09:41 | |
| *** zqfan has joined #openstack-ceilometer | 09:41 | |
| rainmantea | ildokov: well my cloud nodes are CentOS boxes... on a physical node running vm ware | 09:58 |
| rainmantea | on top of these CentOS boxes runs the Openstack cloud | 09:59 |
| *** asalkeld has left #openstack-ceilometer | 10:01 | |
| *** zqfan has quit IRC | 10:08 | |
| *** zqfan has joined #openstack-ceilometer | 10:09 | |
| *** Guest28381 has quit IRC | 10:20 | |
| *** nadya_ has joined #openstack-ceilometer | 10:29 | |
| *** nadya_ has quit IRC | 10:38 | |
| openstackgerrit | mizeng proposed openstack/ceilometer: fix for https://bugs.launchpad.net/ceilometer/+bug/1396473 Rely on VM uuid (which is unique ID and immutable for VM resource) rather than instance name to fetch system metrics via libvert’s lookupByUUIDString API. https://review.openstack.org/137308 | 10:41 |
| openstackgerrit | Sylvain Afchain proposed openstack/ceilometer: Fix Opencontrail pollster according the API changes https://review.openstack.org/104913 | 10:42 |
| ildikov | rainmantea: a-ha, ok, got it | 10:45 |
| openstackgerrit | Mehdi Abaakouk proposed openstack/ceilometer: Add some rally scenarios https://review.openstack.org/132649 | 10:45 |
| openstackgerrit | mizeng proposed openstack/ceilometer: fix for https://bugs.launchpad.net/ceilometer/+bug/1396473 https://review.openstack.org/137308 | 10:52 |
| jd__ | gentux: still working on aggregate method selection on Gnocchi? | 11:00 |
| *** junhongl has quit IRC | 11:01 | |
| *** junhongl has joined #openstack-ceilometer | 11:02 | |
| *** nadya_ has joined #openstack-ceilometer | 11:04 | |
| gentux | jd__: I didn't had time these last few days :/ | 11:05 |
| gentux | jd__: still on it yes | 11:05 |
| *** ildikov has quit IRC | 11:05 | |
| *** exploreshaifali has quit IRC | 11:20 | |
| *** deepthi has quit IRC | 11:22 | |
| *** claudiub has joined #openstack-ceilometer | 11:24 | |
| *** renatoarmani has joined #openstack-ceilometer | 11:28 | |
| openstackgerrit | Ilya Tyaptin proposed openstack/ceilometer: Add encoding of rows and qualifiers in impl_hbase https://review.openstack.org/124380 | 11:36 |
| eglynn | EmilienM: hey, quick puppet-ceilometer merge policy question? | 11:44 |
| eglynn | EmilienM: ... any idea why this puppet-ceilometer patch hasn't merged https://review.openstack.org/134356 despite being +2/+A'd? | 11:44 |
| eglynn | I can't see a related stuck build in zuul | 11:45 |
| * eglynn tries a "reverify" ... | 11:46 | |
| *** ildikov has joined #openstack-ceilometer | 12:11 | |
| *** rainmantea has quit IRC | 12:36 | |
| *** exploreshaifali has joined #openstack-ceilometer | 13:04 | |
| EmilienM | eglynn: checking | 13:11 |
| EmilienM | (good morning) | 13:11 |
| *** Longgeek_ has quit IRC | 13:11 | |
| eglynn | EmilienM: good morning, thanks! | 13:11 |
| eglynn | EmilienM: (I manually triggered a reverify, but still not merged, I may be missing something obvious) | 13:12 |
| EmilienM | eglynn: you did not miss something | 13:17 |
| EmilienM | I'll figure that out | 13:17 |
| eglynn | EmilienM: thank you sir! | 13:17 |
| *** julim has quit IRC | 13:19 | |
| EmilienM | eglynn: I'm gonna +2 +A to try | 13:19 |
| eglynn | cool | 13:20 |
| *** nosnos has quit IRC | 13:20 | |
| *** alexpilotti has joined #openstack-ceilometer | 13:21 | |
| EmilienM | eglynn: looking at http://status.openstack.org/zuul/ it's in the gate now. Should be merged in a few minutes. | 13:22 |
| eglynn | EmilienM: nice one, excellent :) | 13:22 |
| EmilienM | merged | 13:27 |
| *** alexpilotti has quit IRC | 13:32 | |
| *** yatin has quit IRC | 13:32 | |
| *** safchain has quit IRC | 13:33 | |
| *** safchain has joined #openstack-ceilometer | 13:33 | |
| *** jaypipes has joined #openstack-ceilometer | 13:36 | |
| *** ildikov has quit IRC | 13:36 | |
| *** Longgeek has joined #openstack-ceilometer | 13:37 | |
| *** ryanpetrello has joined #openstack-ceilometer | 13:42 | |
| *** gordc has joined #openstack-ceilometer | 13:42 | |
| *** ildikov has joined #openstack-ceilometer | 13:48 | |
| *** julim has joined #openstack-ceilometer | 13:49 | |
| *** nadya_ has quit IRC | 13:54 | |
| openstackgerrit | gordon chung proposed openstack/ceilometer: modify events sql schema to reduce empty columns https://review.openstack.org/130869 | 13:54 |
| openstackgerrit | Julien Danjou proposed stackforge/gnocchi: rest: add and expose back_window attribute of archive policies https://review.openstack.org/136112 | 13:57 |
| *** IvanBerezovskiy has left #openstack-ceilometer | 14:02 | |
| openstackgerrit | Merged openstack/ceilometer: Database.max_retries only override on sqlalchemy side https://review.openstack.org/136964 | 14:14 |
| *** exploreshaifali has quit IRC | 14:15 | |
| openstackgerrit | gordon chung proposed openstack/ceilometer-specs: add ElasticSearch driver backend for events https://review.openstack.org/126395 | 14:17 |
| *** Longgeek has quit IRC | 14:34 | |
| *** nadya_ has joined #openstack-ceilometer | 14:34 | |
| *** r-daneel has joined #openstack-ceilometer | 14:37 | |
| openstackgerrit | Merged openstack/ceilometer: Add encoding of rows and qualifiers in impl_hbase https://review.openstack.org/124380 | 14:38 |
| *** Longgeek has joined #openstack-ceilometer | 14:40 | |
| openstackgerrit | Dina Belova proposed openstack/ceilometer: Move central agent code to the polling agent module https://review.openstack.org/127186 | 14:58 |
| openstackgerrit | Dina Belova proposed openstack/ceilometer: Make compute discovery pollster-based, not agent-level https://review.openstack.org/127185 | 14:58 |
| openstackgerrit | Dina Belova proposed openstack/ceilometer: ==POC== Merge Central and Compute agents to *polling agent* https://review.openstack.org/124719 | 14:58 |
| openstackgerrit | Dina Belova proposed openstack/ceilometer: Merge Central and Compute agents to *polling agent* https://review.openstack.org/124719 | 15:01 |
| *** Kennan has quit IRC | 15:09 | |
| *** ildikov has quit IRC | 15:11 | |
| *** exploreshaifali has joined #openstack-ceilometer | 15:19 | |
| *** Titilambert has quit IRC | 15:23 | |
| *** Titilambert has joined #openstack-ceilometer | 15:24 | |
| *** ildikov has joined #openstack-ceilometer | 15:24 | |
| openstackgerrit | Merged openstack/ceilometer: Internal error with period overflow https://review.openstack.org/134415 | 15:25 |
| openstackgerrit | Dina Belova proposed openstack/ceilometer: Merge Central and Compute agents to *polling agent* https://review.openstack.org/124719 | 15:34 |
| *** ddieterly has joined #openstack-ceilometer | 15:36 | |
| *** alexpilotti has joined #openstack-ceilometer | 15:39 | |
| *** fnaval has joined #openstack-ceilometer | 15:40 | |
| openstackgerrit | Mehdi Abaakouk proposed stackforge/gnocchi: Allows to filter out the gnocchi generated samples https://review.openstack.org/128922 | 15:42 |
| openstackgerrit | Mehdi Abaakouk proposed stackforge/gnocchi: Add a gnocchi dispatcher for ceilometer https://review.openstack.org/98798 | 15:42 |
| *** alla__ has quit IRC | 15:53 | |
| *** ryanpetrello has quit IRC | 16:06 | |
| *** _cjones_ has joined #openstack-ceilometer | 16:08 | |
| *** ishant has quit IRC | 16:09 | |
| *** zigo has quit IRC | 16:16 | |
| *** ildikov has quit IRC | 16:24 | |
| *** zigo has joined #openstack-ceilometer | 16:27 | |
| *** nellysmitt has quit IRC | 16:30 | |
| *** k4n0 has quit IRC | 16:30 | |
| *** ryanpetrello has joined #openstack-ceilometer | 16:31 | |
| *** packet has joined #openstack-ceilometer | 16:35 | |
| *** packet has quit IRC | 16:37 | |
| *** eglynn is now known as eglynn-afk | 16:38 | |
| *** amalagon has quit IRC | 16:44 | |
| *** ryanpetrello_ has joined #openstack-ceilometer | 16:45 | |
| *** ryanpetrello has quit IRC | 16:48 | |
| *** ryanpetrello_ is now known as ryanpetrello | 16:48 | |
| *** ildikov has joined #openstack-ceilometer | 16:53 | |
| *** _cjones_ has quit IRC | 16:54 | |
| *** changbl has joined #openstack-ceilometer | 16:56 | |
| openstackgerrit | Fabio Giannetti proposed openstack/ceilometer: RBAC Support for Ceilometer API Implementation https://review.openstack.org/115717 | 16:58 |
| *** rbak has joined #openstack-ceilometer | 17:06 | |
| *** ryanpetrello has quit IRC | 17:12 | |
| *** Longgeek has quit IRC | 17:24 | |
| *** eglynn-afk is now known as eglynn | 17:25 | |
| *** atan8 has joined #openstack-ceilometer | 17:27 | |
| *** zul has quit IRC | 17:34 | |
| *** zul has joined #openstack-ceilometer | 17:35 | |
| *** ryanpetrello has joined #openstack-ceilometer | 17:35 | |
| *** _cjones_ has joined #openstack-ceilometer | 17:37 | |
| *** _cjones_ has quit IRC | 17:41 | |
| *** _cjones_ has joined #openstack-ceilometer | 17:41 | |
| *** amalagon has joined #openstack-ceilometer | 17:55 | |
| *** sbfox has joined #openstack-ceilometer | 17:58 | |
| *** claudiub has quit IRC | 17:59 | |
| *** amalagon has quit IRC | 18:00 | |
| *** nadya_ has quit IRC | 18:00 | |
| *** ryanpetrello_ has joined #openstack-ceilometer | 18:01 | |
| *** safchain has quit IRC | 18:04 | |
| *** harlowja_away is now known as harlowja | 18:04 | |
| *** ryanpetrello has quit IRC | 18:04 | |
| *** ryanpetrello_ is now known as ryanpetrello | 18:04 | |
| *** _cjones_ has quit IRC | 18:07 | |
| *** _cjones_ has joined #openstack-ceilometer | 18:08 | |
| *** exploreshaifali has quit IRC | 18:13 | |
| *** harlowja has quit IRC | 18:18 | |
| *** harlowja has joined #openstack-ceilometer | 18:19 | |
| *** ryanpetrello_ has joined #openstack-ceilometer | 18:30 | |
| *** nadya_ has joined #openstack-ceilometer | 18:30 | |
| *** ryanpetrello has quit IRC | 18:33 | |
| *** ryanpetrello_ is now known as ryanpetrello | 18:33 | |
| *** pradk has joined #openstack-ceilometer | 18:35 | |
| *** nadya_ has quit IRC | 18:39 | |
| *** harlowja_ has joined #openstack-ceilometer | 18:41 | |
| *** changbl has quit IRC | 18:44 | |
| *** harlowja has quit IRC | 18:45 | |
| *** sbfox has quit IRC | 18:49 | |
| *** sbfox has joined #openstack-ceilometer | 18:51 | |
| sbfox | Hey Ceilometer folks, is anyone aware of an open sourced billing project designed with ceilometer in mind? | 18:53 |
| *** amalagon has joined #openstack-ceilometer | 19:09 | |
| *** zqfan has quit IRC | 19:13 | |
| *** zqfan has joined #openstack-ceilometer | 19:14 | |
| *** ddieterly has quit IRC | 19:21 | |
| *** changbl has joined #openstack-ceilometer | 19:29 | |
| *** renatoarmani has quit IRC | 19:30 | |
| *** ddieterly has joined #openstack-ceilometer | 19:34 | |
| *** edmondsw has joined #openstack-ceilometer | 19:36 | |
| *** exploreshaifali has joined #openstack-ceilometer | 19:37 | |
| edmondsw | gordc, have a few minutes to chat about https://review.openstack.org/#/c/132097/ ? | 19:38 |
| *** amalagon has quit IRC | 19:47 | |
| gordc | edmondsw: sorry, was getting help setting something up... i have a few minutes now. | 19:55 |
| *** rbak_ has joined #openstack-ceilometer | 19:55 | |
| *** rbak has quit IRC | 19:56 | |
| edmondsw | gordc, did you see my latest comment there, or the comments I put in https://review.openstack.org/#/c/115717/ ? | 19:57 |
| edmondsw | thought it might be better to talk through that on IRC than via review comments | 19:58 |
| *** ryanpetrello has quit IRC | 20:00 | |
| gordc | edmondsw: just read it. i should think the RBAC work would cover this... is this a gap in Fabio's design? | 20:01 |
| edmondsw | gordc, at least partially... but as I've dug into it, I think the problem goes deeper than that | 20:02 |
| gordc | tbh, it looks really strange having http.request and http.response hardcoded everywhere. | 20:02 |
| edmondsw | yeah, and that's not the extent of audit data, either... | 20:02 |
| gordc | especially since there isn't audit data in those meters all the time (only if you use pycadf audit middleware) | 20:02 |
| *** ryanpetrello has joined #openstack-ceilometer | 20:02 | |
| edmondsw | that's a discussion that is going to affect any solution, though... something is going to have to distinguish what is audit data and what isn't | 20:03 |
| edmondsw | that's part of the problem... ceilometer doesn't distinguish that today... and it really needs to | 20:04 |
| edmondsw | audit data should not be lumped together with non-audit data that has different TTL requirements, different security requirements, etc. | 20:04 |
| *** nadya_ has joined #openstack-ceilometer | 20:04 | |
| edmondsw | s/TTL/retention/ | 20:04 |
| gordc | agreed. just to clarify, the RBAC solution can't cover it or doesn't cover it? | 20:04 |
| edmondsw | probably both | 20:05 |
| edmondsw | at least doesn't | 20:05 |
| edmondsw | the middleware that I wrote as an alternative has to do a lot of crazy things to try to strip sensitive audit data out of places you wouldn't expect it to have been in the first plce | 20:06 |
| gordc | i see... so i'm actually working on events and the http.* meters will probably end up going there in future... | 20:06 |
| edmondsw | tbf, the changeset I have in https://review.openstack.org/#/c/132097/ doesn't go far enough... I'd have to expand on that to match the additional things I found while working on the middleware approach | 20:07 |
| gordc | what would be a good way to handle restriction (without hardcoding http.* into api | 20:07 |
| edmondsw | let's talk about the things I found so you understand the full picture | 20:07 |
| gordc | i would assume different deployers would have different metrics which they would deem to be sensitive and not sensitive | 20:08 |
| gordc | ok cool | 20:08 |
| edmondsw | I should paste the code so you can see what I'm talking about... one sec | 20:09 |
| gordc | sounds good | 20:09 |
| *** ryanpetrello has quit IRC | 20:10 | |
| edmondsw | gordc, http://paste.openstack.org/show/139034/ | 20:12 |
| edmondsw | so first, I had to add query parms specifying project id for all users (probably only necessary for admins, since non-admins were already restricted to their project, but better to be safe) and user_id for non-admin users | 20:14 |
| edmondsw | being careful to account for the possibility of the user already specifying project/user in query parms | 20:15 |
| edmondsw | then around line 138 you start to see the other things we have to account for | 20:16 |
| gordc | what rules are we trying to enforce: if admin, see everything and only audit data for your project. if not admin, see only project and no audit data? | 20:18 |
| edmondsw | there are several things, starting there, that you can only filter out of responses rather than add query parms to requests for | 20:19 |
| edmondsw | yes | 20:19 |
| *** ryanpetrello has joined #openstack-ceilometer | 20:20 | |
| edmondsw | the /v2/meters response is one interesting case... besides telling you what meters there are, it tries to tell you the project and user for a meter... maybe that makes sense for some meters, but for http.request/response there isn't really a user/project | 20:20 |
| gordc | i guess if we added a flag to audit data, it'd make everything easier? | 20:21 |
| edmondsw | the user/project for http.request/response appears to be returned as the last user/project to access that API, which is a) a bad design and b) sensitive | 20:21 |
| gordc | yeah. there are a lot of 'meters' in ceilometer that aren't meters but events... http.* meters being two of them | 20:21 |
| gordc | http.* meters are what they are because the events part of ceilometer never got completed | 20:22 |
| edmondsw | resources are even worse... when you try to query the resources, you see the full sample data for the last sample on that resource... so I had to trim that off as well | 20:22 |
| gordc | or that's my understanding of it... because outside of the metadata, the top-level attributes make no sense. | 20:22 |
| edmondsw | and it might be possible to edit complex queries during the request phase (yuck! have fun with that...), but I decided to just parse them out of the response instead... still yuck | 20:23 |
| gordc | i'm wondering if this can be filtered out using the complex query stuff. (i assume your current patch doesn't do this 'trim'ing) | 20:24 |
| gordc | ok. | 20:24 |
| edmondsw | I'm not a complex query expert... didn't even know they existed until I was digging in the v2 controllers api code and saw them | 20:24 |
| edmondsw | these APIs should really be documented... | 20:24 |
| gordc | the complex stuff is... i guess not well enough | 20:25 |
| edmondsw | http://developer.openstack.org/api-ref-telemetry-v2.html seems woefully lacking | 20:25 |
| edmondsw | doesn't mention queries at all | 20:25 |
| edmondsw | or /v2/samples | 20:26 |
| edmondsw | or events | 20:26 |
| edmondsw | etc. | 20:26 |
| gordc | i can't speak for those docs... i think we just started on them in juno... | 20:26 |
| gordc | main docs are here: http://docs.openstack.org/developer/ceilometer | 20:27 |
| gordc | http://docs.openstack.org/developer/ceilometer/webapi/v2.html#complex-query | 20:27 |
| edmondsw | tx, I'll look at those | 20:27 |
| *** _cjones_ has quit IRC | 20:28 | |
| edmondsw | you think all this is fixing with moving audit data to events, or will we have some of the same issues there? | 20:30 |
| gordc | same issues. i just want to make sure i address them while i'm working on it. | 20:30 |
| gordc | so the RBAC patch only covers which calls can be made... but doesn't filter out audit data (when appropriate) | 20:32 |
| edmondsw | while you're at it, can we trim down the data that's stored to just the CADF info? It's a waste using up disk storing the service catalog, etc. for each request/response | 20:32 |
| gordc | edmondsw: yeah, there's a event_defintion file which parses out only the attributes you want/need | 20:33 |
| edmondsw | right, the RBAC patch only supports rules based on the requester's role | 20:33 |
| gordc | that technically exists in Juno already so you could theoretically just switch to use events instead of meters... still have same access issues probably. | 20:34 |
| *** _cjones_ has joined #openstack-ceilometer | 20:35 | |
| *** hhuang has quit IRC | 20:35 | |
| *** hhuang has joined #openstack-ceilometer | 20:36 | |
| edmondsw | and I'm not sure how you'd fashion a policy.json rule to say "admins can view everything in the project, non-admins can only view what they own", much less say add "when it comes to audit data" to that | 20:36 |
| edmondsw | oh, and restrict POST of samples to non-audit data | 20:37 |
| *** exploreshaifali has quit IRC | 20:37 | |
| edmondsw | so you don't have people faking audit entries | 20:37 |
| edmondsw | I think we should really separate audit data from other things... /v2/audit ? | 20:38 |
| edmondsw | maybe even a completely new service separate from ceilometer... | 20:38 |
| gordc | how do you know it's audit data? | 20:39 |
| *** nadya_ has quit IRC | 20:39 | |
| gordc | not sure why you'd need a new service or what it'd do. | 20:39 |
| edmondsw | whoever creates the data should know whether it's audit data or not | 20:41 |
| edmondsw | if we can solve the problems within ceilometer that's obviously fine | 20:42 |
| edmondsw | but let's say we move audit info into events... how do we avoid the same RBAC issues there? | 20:43 |
| edmondsw | someone who wants to read audit info is only going to want audit info. Someone who wants to read other events is only going to want to read other events. | 20:43 |
| edmondsw | Nobody is going to want to read both audit info and non-audit event info at the same time | 20:44 |
| edmondsw | so why lump them together under the same API? | 20:44 |
| edmondsw | especially when you need to setup different RBAC rules for accessing audit info than for non-audit info | 20:45 |
| edmondsw | it'd be easier to have them as separate APIs, which can then easily have separate associated policy enforcement actions | 20:45 |
| edmondsw | there will be other differences between audit and non-audit info as well... e.g. retention policies for audit info will need to be distinct from those for non-audit info | 20:46 |
| gordc | i don't think conditional rententoin policies is that difficult. the main thing is to have a way to let ceilometer know 'this is audit data'... | 20:52 |
| gordc | right now what we get is 'this is data' | 20:52 |
| *** zqfan has quit IRC | 20:52 | |
| gordc | and a live person saying 'actually this, this and this is audit data' | 20:53 |
| *** zqfan has joined #openstack-ceilometer | 20:53 | |
| edmondsw | gordc, so we need to change that, definitely... | 20:55 |
| *** nadya_ has joined #openstack-ceilometer | 20:56 | |
| edmondsw | I can't claim to understand how all the audit data comes in well enough to propose a specific answer there... | 20:57 |
| *** amalagon has joined #openstack-ceilometer | 20:57 | |
| edmondsw | but in general, the folks supplying the data should know it's audit data, so they should be able to tell ceilometer that if ceilometer comes up with a way for them to do that | 20:58 |
| gordc | edmondsw: so when i created the new audit middleware in keystonemiddleware, i prepended audit.* to the event_type... https://review.openstack.org/#/c/102958/21/keystonemiddleware/audit.py | 20:58 |
| *** fnaval has quit IRC | 20:58 | |
| gordc | in that case, the events api can easily filter out those events i would think if you're not admin | 20:59 |
| gordc | although i think by default right now, events requires you to be admin regardless to access data. | 21:00 |
| *** amalagon has quit IRC | 21:02 | |
| edmondsw | and events don't have this nonsense about an admin in project A being able to see data in project B, I hope? | 21:02 |
| edmondsw | if it's ok to restrict all events to admins, that might be ok... if someone needs a non-admin to have project-wide access to events, though, we're going to run into trouble unless we enforce audit access separately | 21:03 |
| edmondsw | what about audit data other than http.request/response? | 21:05 |
| edmondsw | not sure we can change them all to prepend "audit." | 21:05 |
| *** atan8 has quit IRC | 21:05 | |
| gordc | i need to check api... i think it's all or nothing. | 21:05 |
| edmondsw | may need to keep a list that ceilometer checks against | 21:05 |
| edmondsw | I'm gonna have to run, but there's some food for thought... let's catch up again after Thanksgiving | 21:06 |
| gordc | that is fine... event_type is a key attribute of events so i think a list could work too. | 21:07 |
| gordc | i'll think about it as i go along or you can play with events and let me know | 21:07 |
| gordc | have a good thanksgiving (even though it was over a month ago) | 21:07 |
| gordc | ;) | 21:08 |
| *** MasterPiece has joined #openstack-ceilometer | 21:08 | |
| edmondsw | :P thanks | 21:08 |
| *** edmondsw has quit IRC | 21:09 | |
| *** atan8 has joined #openstack-ceilometer | 21:10 | |
| *** MasterPiece has quit IRC | 21:10 | |
| *** MasterPiece has joined #openstack-ceilometer | 21:10 | |
| *** MasterPiece has quit IRC | 21:11 | |
| *** MasterPiece has joined #openstack-ceilometer | 21:12 | |
| *** zqfan has quit IRC | 21:13 | |
| *** MasterPiece has quit IRC | 21:13 | |
| *** zqfan has joined #openstack-ceilometer | 21:13 | |
| *** fnaval has joined #openstack-ceilometer | 21:15 | |
| *** pradk has quit IRC | 21:23 | |
| *** atan8 has quit IRC | 21:24 | |
| *** nadya_ has quit IRC | 21:26 | |
| *** atan8 has joined #openstack-ceilometer | 21:27 | |
| *** ryanpetrello has quit IRC | 21:32 | |
| openstackgerrit | litong01 proposed openstack/ceilometer: add http dispatcher https://review.openstack.org/109853 | 21:36 |
| *** amalagon has joined #openstack-ceilometer | 21:50 | |
| *** eglynn_ has joined #openstack-ceilometer | 21:51 | |
| *** eglynn has quit IRC | 21:53 | |
| *** asalkeld has joined #openstack-ceilometer | 22:03 | |
| *** alexpilotti has quit IRC | 22:15 | |
| *** ccrouch has quit IRC | 22:21 | |
| *** promulo has joined #openstack-ceilometer | 22:30 | |
| *** atan8 has quit IRC | 22:35 | |
| *** alexpilotti has joined #openstack-ceilometer | 22:45 | |
| *** EmilienM has quit IRC | 22:47 | |
| *** EmilienM has joined #openstack-ceilometer | 22:47 | |
| *** gordc has quit IRC | 22:55 | |
| openstackgerrit | gordon chung proposed openstack/ceilometer: notification coordination WIP https://review.openstack.org/135414 | 22:58 |
| *** sbfox has quit IRC | 23:01 | |
| *** ddieterly has quit IRC | 23:05 | |
| *** sbfox has joined #openstack-ceilometer | 23:06 | |
| *** ddieterly has joined #openstack-ceilometer | 23:06 | |
| *** zqfan has quit IRC | 23:11 | |
| *** ddieterly has quit IRC | 23:11 | |
| *** zqfan has joined #openstack-ceilometer | 23:11 | |
| *** ryanpetrello has joined #openstack-ceilometer | 23:12 | |
| openstackgerrit | ZhiQiang Fan proposed openstack/python-ceilometerclient: Support ceilometer-url and os-endpoint https://review.openstack.org/137486 | 23:15 |
| *** ryanpetrello has quit IRC | 23:19 | |
| *** nicknach has quit IRC | 23:29 | |
| *** nicknach has joined #openstack-ceilometer | 23:29 | |
| openstackgerrit | ZhiQiang Fan proposed openstack/python-ceilometerclient: Support os-endpoint-type https://review.openstack.org/137488 | 23:31 |
| *** nicknach has left #openstack-ceilometer | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!