15:00:57 <rhochmuth> #startmeeting monasca
15:00:59 <openstack> Meeting started Wed Dec  7 15:00:57 2016 UTC and is due to finish in 60 minutes.  The chair is rhochmuth. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:00 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:02 <openstack> The meeting name has been set to 'monasca'
15:01:03 <rhochmuth> o/
15:01:07 <kamil> o/
15:01:10 <koji> o/
15:01:10 <rbak> o/
15:01:10 <witek_> hello
15:01:13 <rhochmuth> https://etherpad.openstack.org/p/monasca-team-meeting-agenda
15:01:26 <rhochmuth> Agenda for Wednesday December 7 2016 (15:00 UTC)
15:01:27 <rhochmuth> 1.	Grafana branch - merge upstream?
15:01:27 <rhochmuth> 2.	Reviews:
15:01:27 <rhochmuth> 1.	https://review.openstack.org/#/c/395246/
15:01:27 <rhochmuth> 2.	https://review.openstack.org/#/c/349097 - SAP team would like to contribute
15:01:27 <rhochmuth> 3.	https://review.openstack.org/406936, https://review.openstack.org/407371
15:01:27 <rhochmuth> 4.	https://review.openstack.org/407379
15:01:27 <rhochmuth> 5.	https://review.openstack.org/#/c/395897/
15:01:33 <rhochmuth> hi everyone
15:01:41 <hosanai> o/
15:01:49 <rhochmuth> it is really cold and snowing here
15:02:00 <qwebirc30600> 0/
15:02:14 <qwebirc30600> where?
15:02:16 <witek_> nice
15:02:24 <kamil> fridge opened?
15:02:28 <rhochmuth> Fort Collins, CO
15:02:42 <qwebirc30600> o
15:02:53 <rhochmuth> let's get started with the agenda first
15:03:05 <rhochmuth> #topic Grafana branch - merge upstream?
15:03:09 <rbak> That's me
15:03:21 <rhochmuth> please proceed
15:03:42 <rbak> Raintank, the owner's of Grafana have been fairly silent lately
15:03:52 <rhochmuth> uhh, ohhh
15:04:25 <rbak> We're not sure that our original plan of brining them on as vendors for Charter is going to work
15:04:48 <rbak> And our fork of Grafana is slowing getting more complex and harder to merge in.
15:04:59 <rhochmuth> do you think that they are having financial problems
15:05:06 <rhochmuth> or just not interested in what we are doing
15:05:14 <rhochmuth> or other priorities
15:05:15 <rbak> Just not interested
15:05:27 <rbak> That's probably more accurate
15:05:52 <rbak> So we either need to make a push as a community to get the keystone changes moved into upstream, or we need to resign ourselves to maintaining a permanent fork again.
15:05:52 <rhochmuth> i thought you guys were going to put a support agreement or licensing in-place
15:06:05 <rhochmuth> that would have incenticized them enough
15:06:17 <rbak> We're still trying, but they stopped replying.
15:06:26 <rbak> The difficulties with the merger may have scared them off
15:06:44 <rhochmuth> i see
15:07:18 <rbak> So it's a question of whether we can rally enough support
15:07:30 <rbak> I can put up a new pull request
15:07:40 <rhochmuth> ok
15:07:51 <rhochmuth> are we just supposed to thumbs up it at that point
15:08:00 <rhochmuth> what does support mean from your viewpoint?
15:08:08 <rbak> Thumbs up would be good
15:08:14 <rbak> I also plan on emailing them directly.
15:08:15 <rhochmuth> ok, we can do that
15:08:21 <rhochmuth> ok
15:08:26 <rbak> If anyone wants to be included on that let me know
15:08:26 <rhochmuth> i can contact them also
15:08:34 <rhochmuth> please include me
15:08:45 <rbak> will do
15:08:46 <rhochmuth> are you talking to raj dutt
15:08:50 <rhochmuth> i think that is the name
15:08:57 <rhochmuth> it has been a while
15:09:09 <rbak> Yeah, him and Torkel if I can dig out his email.
15:09:34 <rhochmuth> with more and more folks using monasca, i'm hoping they would be interested
15:09:48 <rhochmuth> there are potential support/service agreements they could get as a result
15:10:16 <rbak> That and the fact that the changes we want are actually keystone specific, so in theory it could be used with any openstack projects, or just Grafana as a service.
15:10:26 <rhochmuth> but, there is a lot of momentum around some other tools, like promethues and influxdb, that is probably distractign them
15:11:06 <rbak> Well, they just released Grafana 4, so hopefully this is a good time, while they're still planning for the next version.
15:11:11 <rhochmuth> yes, i agree, i would expect some interest on their part
15:11:36 <rhochmuth> so, now that 4 is out, are you going to look at adding alerts?
15:11:39 <rhochmuth> :-)
15:11:48 <rbak> We're going to look into it, yes
15:11:53 <rhochmuth> thanks
15:12:11 <rhochmuth> we've been using grafana a lot internally lately, and it is working great
15:12:17 <rbak> Good to hear
15:12:28 <rhochmuth> we had some very impressive demos this week with kubernetes
15:12:34 <rhochmuth> and monasca
15:12:49 <rbak> That's all I had on this subject.  I'll put up a new pull request and bring that up at the meeting next week.
15:13:26 <rhochmuth> i'm wondering if we should get a higher level business call together with raintank between allthe interested companies
15:13:37 <rhochmuth> charter, hpe, fujitsu, …,
15:13:41 <rbak> I would be up for that.
15:13:48 <rbak> I'll suggest that in my email
15:13:56 <witek_> you have our support too
15:13:56 <rhochmuth> yes, please do
15:14:15 <rhochmuth> i think they should understand the momentum and business opportunity better
15:14:29 <rhochmuth> maybe we should get and email out to openstack-dev list as well
15:14:47 <rhochmuth> there are other companies involved in deploying monasca these days that we could also lobby
15:15:05 <rhochmuth> invariably, it is going to be about the money i guess
15:15:19 <rhochmuth> or the perceived opportunity
15:15:51 <rbak> Probably, but they are open source, and did tell us they were interested even without a contract when we talk initially.
15:16:19 <rhochmuth> ok, let's see what happens with the email
15:16:28 <rbak> Sounds good.  I'll include you on that.
15:16:30 <rhochmuth> and then maybe we can rally the troops
15:16:41 <rhochmuth> thanks rbak
15:16:44 <rhochmuth> good update
15:16:59 <rhochmuth> #topic https://review.openstack.org/#/c/395246/
15:17:13 <rbak> That one is also mine
15:17:25 <rbak> Just wondering if it can be merged
15:17:35 <rbak> It's been around a couple weeks with some +1s
15:17:36 <rhochmuth> i +1'd again this morning, but yes i think it is ready
15:18:05 <rhochmuth> you should probably ping craig
15:18:13 <rhochmuth> as he spotted a couple of issues
15:18:27 <rbak> Alright, I can get him to look at it
15:18:31 <rhochmuth> but he hasn't looked at it again since your latest fixes/reviews
15:19:03 <rhochmuth> assuming he is ok, then it could be merged
15:19:14 <rbak> Sounds good.  I'll ping him
15:19:29 <rhochmuth> thx
15:19:33 <rhochmuth> #topic https://review.openstack.org/#/c/349097
15:19:56 <rhochmuth> sap, jbors u there?
15:20:03 <jobrs> yep
15:20:11 <jobrs> we are willing to help.
15:20:17 <rhochmuth> thanks
15:20:20 <jobrs> and we are working on some complementary stuff
15:20:33 <rhochmuth> so, do you want to take on this review
15:20:42 <rhochmuth> i don't think haneef is working on it anymore
15:20:48 <rhochmuth> and was just waiting to merge it
15:20:57 <rhochmuth> i had signed off on it
15:21:01 <jobrs> sure
15:21:06 <rhochmuth> thanks
15:21:24 <dhague> sorry, late to the topic and just catching up on the discussion so far - all sounds good to me. We need to get some kind of response from raintank, even if it's jsut to get an answer one way or the other
15:21:24 <jobrs> we want to add notification templates for slack and mail since this is what we are using
15:21:47 <rhochmuth> dhague: yes, i agree
15:21:56 <dhague> ... and I just noticed the topic change. sorry 'bout that
15:22:10 <rhochmuth> rbak: please add dhague and jbors to email list, i'm assuming
15:22:24 <rbak> will do
15:22:25 <rhochmuth> email list to raintank that is
15:22:27 <jobrs> thanks
15:22:33 <dhague> +1
15:22:35 <rhochmuth> ok, np on topic change
15:23:32 <rhochmuth> so, i'm assuming dhague will take over the custom formatting for the hipchat plugin
15:23:49 <rhochmuth> you had also fixed the slack plugin recently too
15:23:52 <dhague> on the new topic, I am looking at that - in fact, having some kind of per-channel templating
15:24:02 <dhague> the idea is this:
15:24:38 <dhague> alarm definition "description" field will be a jinja template, so we can generate some kind of human-readable alert message
15:24:57 <bklei> question on the hipchat plugin formatting
15:25:11 <dhague> then at the channel level that can be fed into another template - pretty raw for hipchat & slack, but email would allow HTML headers etc
15:25:29 <bklei> will this patch let us change msg color based on alarm severity?
15:25:49 <seanhandley> Oh, the public cloud meeting's over
15:25:51 <seanhandley> :(
15:25:55 <jobrs> that will be possible
15:26:06 <jobrs> jinja2 has primitives like if and for etc.
15:26:22 <bklei> excellent -- it's goofy that everything is green today
15:26:49 <jobrs> the redesign has two parts
15:27:00 <dhague> jobrs and I had a brainstorming on this yesterday, the above is the brief summary of what we discussed - is it OK with everyone?
15:27:07 <dhague> I will let jobrs continue...
15:27:26 <jobrs> just wanted to add what I discovered today
15:27:46 <jobrs> so part one is just in notification: templates for formatting notification messages
15:28:49 <jobrs> part two is more tricky: have the monasca-api pass the alarm-definition's description alongside the alarm (not just the name)
15:29:15 <bklei> that's not there already?
15:29:23 <jobrs> so we would need some support in reviewing the API changes and someone who would do the Java side of it
15:29:51 <rhochmuth> why do you need the description in the alarm?
15:30:09 <rhochmuth> can't you get that by querying the db?
15:30:25 <rhochmuth> not sure i understand the proposal
15:30:54 <bklei> having the alarm definition description in the alarm is useful -- we use it for runbook links for oncall
15:31:01 <jobrs> the description is what matters most to the user (imho)
15:31:16 <bklei> but i believe it's already there -- this is from the current hipchat plugin:
15:31:22 <bklei> "old_state": "UNDETERMINED",
15:31:22 <bklei> "alarm_description": "The http response time is greater than 3s on avg",
15:31:22 <bklei> "message": "Thresholds were exceeded for the sub-alarms: avg(soapuiv2MonascaTesterMetric{hostname=dnvrco02-keystone-001}) > 3.0 with the values: [4.6218046456454775]",
15:31:51 <jobrs> it is in the notifications, but not the alarms. when you display the alarm in a UI it would be good to have the description of it
15:32:22 <bklei> aah.  i see.
15:32:58 <rhochmuth> but, that could be done with mutiple queries
15:33:06 <rhochmuth> get alarm, then get alarm definition
15:33:07 <jobrs> if you have a templated alarm description and you can render it using the alarm attributes, then you can have quite instructive alarms
15:33:09 <rhochmuth> then merge the two
15:33:25 <jobrs> slow
15:33:40 <jobrs> you did the job already for the alarm-description-name
15:34:02 <rhochmuth> ok
15:34:15 <rhochmuth> interesting, that we didn't see that through
15:34:44 <rhochmuth> so, when you query an alarm, you would also return the description in the json body?
15:35:03 <jobrs> yes, and I would do the template rendering in the monasca-client
15:36:11 <rhochmuth> so, how would the monasca-client know how to render it?
15:36:34 <rhochmuth> the custom formatting is for the notifications
15:37:19 <jobrs> that is what I mean with two parts. One thing is to have descriptions with variables. the other is to have proper presentation for your notification channel
15:37:37 <jobrs> we believe you need both to have actionable alarms
15:38:50 <rhochmuth> ok
15:39:07 <rhochmuth> not sure i'm grokking the full extent of the changes
15:39:31 <jobrs> it should be non-breaking
15:39:43 <rhochmuth> good enough for me
15:39:51 <rhochmuth> :-)
15:40:09 <dhague> We can discuss the fine details in the code review :-)
15:40:15 <rhochmuth> sounds good
15:40:30 <jobrs> let's do so
15:41:39 <rhochmuth> so, at this point i please proceed, and we can discuss more in the code review
15:41:51 <jobrs> last question - can we use the change for further developments or created a separate one?
15:42:15 <rhochmuth> sure, you can use that review
15:42:26 <rhochmuth> i think that is the best option
15:42:48 <jobrs> great, I hope we can deliver the first proposal next week
15:42:54 <rhochmuth> thanks
15:43:46 <rhochmuth> #topic https://review.openstack.org/406936, https://review.openstack.org/407371
15:44:55 <rhochmuth> yingjun: you there?
15:45:07 <yingjun> yep
15:45:17 <rhochmuth> you are up
15:45:25 <rhochmuth> Fix UnicodeEncodeError for alarm definition
15:46:27 <yingjun> so as the bug reported, when i input Chinese in description, it will raise the UnicodeEncodeError
15:47:42 <rhochmuth> i thought we could encode/decode utf8
15:48:13 <rhochmuth> so, i'm a bit confused by the fix and why it was necessary
15:48:24 <witek_> yingjun: could you add that case to the tests?
15:48:59 <yingjun> witek_, sure
15:49:12 <rhochmuth> shouldn't decode(('utf8') work for both chineese and other strings
15:50:01 <rhochmuth> what you are doing now is not decoding the string if it is six.text_type
15:50:45 <yingjun> yes
15:51:59 <yingjun> i’m not sure if there’s case it wasn’t text_type
15:52:59 <rhochmuth> yes, i'm wondering why the data isnn't stored as text_type
15:53:12 <rhochmuth> i'm wondering if bad data got in the db
15:53:22 <rhochmuth> should we have converted it when storing the description
15:53:35 <rhochmuth> it sounds to me that it wasn't converted when we stored it
15:53:52 <yingjun> the case https://github.com/openstack/monasca-api/blob/master/monasca_api/tests/test_alarms.py#L504 do
15:54:40 <yingjun> but it mocks the db return value
15:55:00 <rhochmuth> ok, i'll look at it more
15:55:18 <rhochmuth> i obviousely am missing some bit of understanding
15:55:33 <rhochmuth> ill add comments to the review
15:55:38 <yingjun> ok
15:55:38 <witek_> have you checked that your db has utf8 encoding?
15:56:06 <yingjun> witek_, not yet
15:56:36 <yingjun> witek_, i will check that tomorrow, it pretty late here in China...
15:57:02 <yingjun> witek_, and the test env not in my laptop
15:57:19 <rhochmuth> #topic https://review.openstack.org/#/c/407379/
15:57:42 <rhochmuth> looks like joe merged that one this morning
15:58:11 <rhochmuth> #topic https://review.openstack.org/#/c/395897/
15:58:21 <rhochmuth> i think we are going to run out of time to discuss this one
15:58:35 <yingjun> ;(
15:58:50 <rhochmuth> i'll try and look at it more
15:59:01 <yingjun> rhochmuth, thanks
15:59:23 <rhochmuth> tomasz seems to think it was the right direction too
15:59:37 <rhochmuth> he should probably be pinged again on this one
15:59:45 <rhochmuth> so, i need to close out the meeting again
15:59:51 <rhochmuth> one hour just isn't enough some days
16:00:04 <rhochmuth> thanks everyone
16:00:11 <dhague> thanks
16:00:17 <qwebirc30600> thanks
16:00:26 <rhochmuth> #endmeeting