08:01:13 #startmeeting vitrage 08:01:13 Meeting started Wed Jul 27 08:01:13 2016 UTC and is due to finish in 60 minutes. The chair is ifat_afek. Information about MeetBot at http://wiki.debian.org/MeetBot. 08:01:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 08:01:16 The meeting name has been set to 'vitrage' 08:01:33 hi 08:01:53 hi 08:02:00 Hi everyone! 08:02:13 :) 08:04:21 Hi 08:04:32 Our agenda today: 08:04:37 * Status and Updates 08:04:42 * Open Discussion 08:04:52 #topic Status and Updates 08:05:16 The voting for Barcelona sessions has started! 08:05:25 #link https://www.openstack.org/summit/barcelona-2016/vote-for-speakers/ 08:05:37 Vitrage submitted 8 sessions 08:05:54 Please enter the link, and select "Vitrage" in the search box to get the first 7 sessions. Then, select "RCA" to get the last one 08:06:02 Good luck to us!! 08:06:12 Vitrage mascot selection - our list has slightly changed. 08:06:24 Heidi Joy Tretheway, who leads the mascot selection process, said we won't be able to vote for the tree as Senlin are using a forest. I asked her to reconsider it 08:06:31 In addition, Telemetry selected a Meerkat, which is a Surricata. I agreed to let them have it. 08:06:41 After we lost our two favorite candidates, we thought of adding a Giraffe. It sees everything from above, and its skin can be colored like a Vitrage 08:06:50 So our candidates would be: Giraffe, Parrot and Butterfly. If you have other suggestions, maybe we can still add them to the list. Heidi is supposed to close the lists today (for everyone ) and then we should vote 08:06:55 #link https://etherpad.openstack.org/p/vitrage-mascot 08:07:05 Who wants to update? 08:07:43 Hi, I'll update 08:08:10 I'm nearly done with the zabbix datasource notifications, hopefully I'll get it pushed by the end of this day. 08:10:04 Along with the vitrage part of the code there's also an auxiliary script written which is needed be attached to zabbix as a mediatype, the script is responsible for sending vitrage notifications per zabbix event i.e. trigger value changed from problem to ok and vice versa 08:10:26 that's it for now.. :) 08:10:41 nbloom: and the script should be configured in zabbix? 08:11:52 yes, along with the script there is also a readme explaining the process of linking the script to zabbix. In addition, there's a similar rst file that will be published with vitrage's docs 08:12:12 nbloom: great, thanks 08:12:27 also, please add a blueprint for zabbix notifications datasoruce 08:12:45 sure, will be done today! 08:12:58 I have an update 08:13:19 I am working on Vitrage compitability for liberty cycle 08:14:22 It is almost working, but still have some problem with the change passing the Vitrage tempests, because the keystoneauth1 version is differernt 08:14:44 it is working in the vitrage code, but I still have a problem in the vitrageclient code 08:14:54 need to change there the authentication 08:15:05 Done :) 08:15:32 alaxey_weyl: thanks! 08:15:54 Anyone else wants to update? 08:17:37 We had one action item from the last meeting: 08:17:53 * danoffek add heat blueprint 08:17:59 He was unable to join the meeting, but I know that he did add this blueprint 08:18:48 I'll go next 08:20:46 As part of Zabbix datasource support, there is also an effort (mainly liat and nofar, and hopefully soon myself as well) to add zabbix-based vitrage-templates 08:21:32 This raises a general issue we need to address: how do we make vitrage work out-of-the-box for users 08:22:08 The dilemma is as follows: vitrage templates use the name of the alarm as part of their structure 08:22:45 but the exact name might change for each user (e.g., a cpu-threshold test can be called "cpu util" "cpu load" etc) 08:23:15 elisha_r: so it depends on the zabbix configuration, right? 08:24:35 I think what we need to do is try and locate *classic* packages for zabbix, nagios, sensu etc., and build template suites for these. That way, people can use Vitrage OOTB by adding these test packages. 08:25:23 elisha_r: I agree. At least a naiive user will be able to use the classic zabbix packages and have vitrage configured without extra effort 08:25:46 and for more complex use cases there will be a need to copy-paste and modify the templates 08:26:49 Even if we do this, though, I think it raises a question of how to store and manage these template suites. We could (1) have them stored as a list of TARs that can be downloaded separately, 08:27:18 why not have them in a folder for copy-paste? 08:27:20 or (2) they could all be installed by default, but then we need to be able to select which is loaded into vitrage, some sort of config file 08:27:48 we could do copy-paste, but I think configuration is more simple to manage, allowing turning on/off easily 08:28:04 regarding [2] - this is in our roadmap. To have a "bank" of templates, and let the user activate or deactivate them. But of course it is much more work to implement 08:28:43 So I guess we can start with [1] and improve it in the next versions 08:29:11 re [2] - I agree, but I think there is room to distinguish between bulk loading and selecting individual tests. 08:29:18 but of course, we should start with [1] 08:30:03 A similar issue should be raised regarding support for external datasource "update" functionality 08:30:19 what do you mean? 08:32:03 for example, for Zabbix support noam (nbloom) wrote a plugin for Zabbix. This is not installed in Vitrage, but in Zabbix, to notify Vitrage of changes. The question is, how to manage these in the code base. For now I think the solution was to have, for each datasource, an "auxiliary" folder to place these, and then have a README/RST to explain where the files should go, how to configure etc. 08:32:40 got it. you are right, we need to decide on a consistent way to handle it 08:33:23 I think you're right too 08:33:43 cool 08:33:48 Any other updates? 08:34:39 #topic Open Discussion 08:34:51 Anything you would like to raise? 08:36:32 I guess we are done for today 08:36:39 Goodbye :-) 08:36:43 bye 08:36:55 bye 08:37:00 #endmeeting