cindyo | sandywalsh: which timestamp in the raw event do you use for the required timestamp field? I see timestamp, context_timestamp, launched_at, created_at. | 00:19 |
---|---|---|
sandywalsh | cindyo: "timestamp" | 00:19 |
cindyo | too obvious :) | 00:19 |
cindyo | thnks | 00:19 |
sandywalsh | launched_at and created_at and deleted_at have specific meanings within NOva | 00:19 |
sandywalsh | *nova | 00:19 |
sandywalsh | and _context_* stuff leaked into notifications a long time ago and were never removed :/ | 00:20 |
cindyo | I’m just getting back to events. I have the topics added to kafka in mini-mon, and now am working on a test driver that feeds events (wrapped in an evelope with tenant id, region which is what we need) to the raw-events topic. Bypassing api for now | 00:21 |
cindyo | using notigen to generate :) | 00:21 |
sandywalsh | heh, nice! | 00:21 |
sandywalsh | we've been tied up with stackforge stuff, but back in the saddle again now | 00:21 |
cindyo | good news! :) | 00:22 |
sandywalsh | I'm moving the oahu stuff into winchester | 00:22 |
sandywalsh | so, we should be rocking some monasca integration soon I hope | 00:22 |
cindyo | yes, please, I’m used to that interface | 00:22 |
sandywalsh | by interface do you mean the Config object vs. the yaml stuff? | 00:22 |
cindyo | I’ll be ready (don’t know about you) next week to start looking at winchester. Is that doable for getting oahu in? | 00:22 |
sandywalsh | I think so ... so long as stackforge doesn't cause any issues | 00:23 |
cindyo | by interface, i mean constructing the pipeline, and creating the nice callback | 00:23 |
sandywalsh | perhaps you, I, roland and dragon should have a little pow-wow tomorrow to sync up? | 00:23 |
sandywalsh | gotcha. yeah, I kinda like the explicit definitions myself too | 00:24 |
cindyo | that would be good | 00:24 |
cindyo | I haven’t really looked at winchester yet to be honest. I heard it has a richer config. So I’ll use it as you recommend of course! | 00:25 |
sandywalsh | they're largely the same, but configuration is better with winchester and I suspect the db driver is more performant | 00:28 |
sandywalsh | I'll send an email about tomorrow | 00:29 |
cindyo | sounds great, looks like we are all free except for a short meeting at 11:00 MDT | 00:30 |
cindyo | anytime except 11:00 | 00:30 |
sandywalsh | k, thanks | 00:30 |
cindyo | oh, and 3-4, rolland has a meeting | 00:31 |
*** openstackgerrit has quit IRC | 00:31 | |
*** openstackgerrit has joined #openstack-monasca | 00:32 | |
*** cindyo has quit IRC | 01:18 | |
*** cindyo has joined #openstack-monasca | 02:56 | |
*** cindyo has quit IRC | 03:43 | |
*** craigbr has joined #openstack-monasca | 04:05 | |
*** garyh has quit IRC | 04:10 | |
*** craigbr has quit IRC | 04:19 | |
*** craigbr has joined #openstack-monasca | 04:23 | |
*** craigbr has quit IRC | 04:25 | |
*** craigbr has joined #openstack-monasca | 04:32 | |
*** craigbr has quit IRC | 04:58 | |
*** sandywalsh has quit IRC | 12:37 | |
*** sandywalsh has joined #openstack-monasca | 12:45 | |
*** craigbr has joined #openstack-monasca | 13:04 | |
*** craigbr has quit IRC | 13:17 | |
*** craigbr has joined #openstack-monasca | 13:29 | |
*** craigbr has quit IRC | 13:56 | |
*** craigbr has joined #openstack-monasca | 13:58 | |
*** craigbr has quit IRC | 14:02 | |
*** craigbr has joined #openstack-monasca | 14:04 | |
*** sandywalsh has quit IRC | 14:09 | |
*** rbak has joined #openstack-monasca | 14:11 | |
*** garyh has joined #openstack-monasca | 14:16 | |
*** bklei has joined #openstack-monasca | 14:19 | |
*** sandywalsh has joined #openstack-monasca | 14:38 | |
*** cindyo has joined #openstack-monasca | 14:41 | |
*** Guest55683 has joined #openstack-monasca | 14:53 | |
*** sandywalsh has quit IRC | 15:02 | |
*** dschroeder has joined #openstack-monasca | 15:04 | |
*** craigbr has quit IRC | 15:10 | |
*** craigbr has joined #openstack-monasca | 15:14 | |
*** tongli has joined #openstack-monasca | 15:14 | |
openstackgerrit | Deklan Dieterly proposed a change to stackforge/monasca-persister: Initial Python Persister https://review.openstack.org/122009 | 15:17 |
*** sandywalsh has joined #openstack-monasca | 15:20 | |
*** craigbr has quit IRC | 15:52 | |
*** sandywalsh has quit IRC | 17:00 | |
*** tongli has quit IRC | 17:00 | |
*** bklei has quit IRC | 17:00 | |
*** sandywalsh has joined #openstack-monasca | 17:00 | |
*** tongli has joined #openstack-monasca | 17:00 | |
*** bklei has joined #openstack-monasca | 17:00 | |
*** jodah has joined #openstack-monasca | 17:30 | |
*** craigbr has joined #openstack-monasca | 17:31 | |
cindyo | sandywalsh: I’m using notigen, and I think one (or more) of the event types does not have ‘tenant_id’. Would that be correct? | 17:50 |
sandywalsh | cindyo: that's possible ... lemme check something | 17:51 |
cindyo | looking for [‘payload’][‘tenant_id’]. Should it always be there?? | 17:51 |
*** Guest55683 is now known as Dragon | 17:52 | |
sandywalsh | cindyo: how about _context_tenant ... is that always there? | 17:52 |
*** Dragon is now known as Guest65780 | 17:52 | |
cindyo | I could check, i only saw the error after processing several events | 17:52 |
sandywalsh | cindyo: payload.tenant_id could be hit or miss ... depends on who fabricated the payload | 17:52 |
sandywalsh | the payload is the Wild West | 17:52 |
sandywalsh | things above payload are a little more consistent | 17:53 |
*** Guest65780 is now known as mdragon | 17:53 | |
cindyo | hmm, maybe i should always test for tenant-id | 17:53 |
sandywalsh | I'd use_context_tenant | 17:53 |
mdragon | yah, payloads can be pretty variable. | 17:53 |
sandywalsh | _context_tenant | 17:54 |
cindyo | thanks, i’ll try that | 17:54 |
cindyo | sandywalsh: would there be a field that consistenly gives something like region? I was using payload availability_zone | 17:56 |
mdragon | cindyo: Alas, most openstack cells/installations are ignorant of region | 17:58 |
sandywalsh | yep, what he said :) | 17:59 |
sandywalsh | it might be good to get that in the winchest or yagi config in some fashion mdragon | 17:59 |
sandywalsh | brb | 17:59 |
cindyo | alas :) | 18:00 |
mdragon | sandywalsh: we actually do in yagi config | 18:00 |
mdragon | that's how rs adds region info to the notifications. Yagi does it. | 18:00 |
mdragon | based on the source queue | 18:00 |
mdragon | but, yah for stv3, it might be usefull to pass some extra traits in when creating the event, so yagi could tell it what it's region is. | 18:02 |
* mdragon adds to winchester TODO list | 18:02 | |
cindyo | we mainly needed tenant_id, and region for the event envelope. I can ask roland if region is absolutiely necessary, but metrics have it | 18:03 |
mdragon | are you processing these notifications through stackdistiller ? | 18:05 |
*** htruta has joined #openstack-monasca | 18:10 | |
openstackgerrit | Jonathan Halterman proposed a change to stackforge/monasca-common: WIP - Event changes for alarmed-metrics support https://review.openstack.org/121652 | 18:17 |
*** bklei has quit IRC | 18:35 | |
*** craigbr has quit IRC | 18:42 | |
*** craigbr has joined #openstack-monasca | 18:47 | |
*** bklei has joined #openstack-monasca | 19:02 | |
*** craigbr has quit IRC | 19:04 | |
openstackgerrit | Deklan Dieterly proposed a change to stackforge/monasca-persister: Initial Python Persister https://review.openstack.org/122009 | 19:06 |
*** cindyo has quit IRC | 19:06 | |
*** cindyo has joined #openstack-monasca | 19:32 | |
*** craigbr has joined #openstack-monasca | 19:36 | |
*** tongli has quit IRC | 19:38 | |
openstackgerrit | David Schroeder proposed a change to stackforge/monasca-agent: Remove 'detail' dimension, fix a couple bugs https://review.openstack.org/122229 | 19:40 |
sandywalsh | interesting microservices article http://www.activestate.com/blog/2014/09/microservices-and-paas-part-iii (cc:mdragon) | 19:42 |
*** jodah has quit IRC | 19:48 | |
*** jodah has joined #openstack-monasca | 19:53 | |
openstackgerrit | Cindy O'Neill proposed a change to stackforge/python-monascaclient: updated README.rst for API clients https://review.openstack.org/121930 | 20:06 |
cindyo | mdragon: yes, but haven’t hooked it up the transform-engine yet | 20:08 |
*** craigbr has quit IRC | 20:16 | |
*** craigbr has joined #openstack-monasca | 20:18 | |
*** jodah has quit IRC | 20:19 | |
*** craigbr has quit IRC | 20:24 | |
mdragon | Ah. yah, just wanted to mention that for things like tenant, you can have distiller look at multiple fields, just make fields a list and it'll grab the first one that's in the notification. So you could look at payload.tenant_id and _context_tenant if there isn't anything in payload | 20:24 |
cindyo | mdragon: that’s cool. I think the reason we want to wrap it at the API in the envelope that has tenant_id is for the tenant custom metric case. The api knows which tenant authenticated | 20:27 |
cindyo | tenant custom event case i mean. Not a nova event | 20:28 |
*** craigbr has joined #openstack-monasca | 20:29 | |
openstackgerrit | David Schroeder proposed a change to stackforge/monasca-agent: Remove 'detail' dimension, fix a couple bugs https://review.openstack.org/122229 | 20:40 |
openstackgerrit | David Schroeder proposed a change to stackforge/monasca-agent: Remove 'detail' dimension, fix a couple bugs https://review.openstack.org/122229 | 20:41 |
mdragon | cindyo: Ah, ok. | 20:53 |
*** craigbr has quit IRC | 20:56 | |
openstackgerrit | A change was merged to stackforge/python-monascaclient: updated README.rst for API clients https://review.openstack.org/121930 | 20:56 |
*** craigbr has joined #openstack-monasca | 21:01 | |
*** craigbr has quit IRC | 21:03 | |
*** craigbr has joined #openstack-monasca | 21:05 | |
cindyo | sandywalsh: is notigen generally faster when you run it in realtime? Since tick is moved for every list of events processed, this would seem slower | 21:20 |
openstackgerrit | David Schroeder proposed a change to stackforge/monasca-agent: Remove 'detail' dimension, fix a couple bugs https://review.openstack.org/122229 | 21:20 |
openstackgerrit | David Schroeder proposed a change to stackforge/monasca-agent: Remove 'detail' dimension, fix a couple bugs https://review.openstack.org/122229 | 21:21 |
*** bklei has quit IRC | 21:24 | |
*** bklei has joined #openstack-monasca | 22:18 | |
*** craigbr has quit IRC | 22:21 | |
cindyo | sandywalsh: I see that realtime is slower, if you asked for 10 per hour. Finally it’s quitting since I gave it an end time limit instead of nevents test | 22:30 |
cindyo | working much better | 22:34 |
*** craigbr has joined #openstack-monasca | 22:48 | |
*** bklei has quit IRC | 22:50 | |
*** craigbr has quit IRC | 22:53 | |
*** bklei has joined #openstack-monasca | 22:54 | |
*** rbak has quit IRC | 22:56 | |
*** dschroeder has quit IRC | 23:05 | |
*** bklei has quit IRC | 23:08 | |
*** bklei has joined #openstack-monasca | 23:08 | |
*** bklei has quit IRC | 23:13 | |
*** jodah has joined #openstack-monasca | 23:37 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!