Wednesday, 2014-06-25

*** prad_ has quit IRC00:09
openstackgerritBoris Pavlovic proposed a change to openstack/ceilometer-specs: Add spec for osprofiler notification plugin
openstackgerritA change was merged to openstack/python-ceilometerclient: Fix alarm-threshold-update --query option
*** jaypipes has quit IRC00:35
*** matsuhashi has joined #openstack-ceilometer00:35
openstackgerritLianhao Lu proposed a change to openstack/ceilometer: Fixed various import issues exposed by unittest
*** ddieterly has joined #openstack-ceilometer00:59
*** _cjones_ has quit IRC01:03
*** _cjones_ has joined #openstack-ceilometer01:04
*** _cjones_ has quit IRC01:08
*** fnaval has joined #openstack-ceilometer01:10
*** Qiming has joined #openstack-ceilometer01:23
*** nati_ueno has quit IRC01:40
*** Qiming has quit IRC01:42
*** ddieterly has quit IRC01:49
*** Ruetobas has quit IRC01:51
*** nosnos has joined #openstack-ceilometer01:51
*** alexpilotti has quit IRC01:55
*** nosnos has quit IRC01:56
*** nosnos has joined #openstack-ceilometer01:57
*** Ruetobas has joined #openstack-ceilometer01:58
*** Ruetobas has quit IRC02:02
*** Ruetobas has joined #openstack-ceilometer02:03
openstackgerritliusheng proposed a change to openstack/ceilometer: Fix SQL exception getting statitics with metaquery
openstackgerritShuangtai Tian proposed a change to openstack/ceilometer: Bump hacking to 0.9.x
*** ddieterly has joined #openstack-ceilometer02:20
*** ddieterly has quit IRC02:25
*** openstackgerrit has quit IRC02:31
*** changbl has joined #openstack-ceilometer02:32
*** ddieterly has joined #openstack-ceilometer02:34
*** ddieterl_ has joined #openstack-ceilometer02:36
*** mitz_ has joined #openstack-ceilometer02:36
*** ddieterly has quit IRC02:38
*** ddieterl_ has quit IRC02:40
*** Alice911 has joined #openstack-ceilometer02:42
*** ddieterly has joined #openstack-ceilometer02:43
*** Qiming has joined #openstack-ceilometer02:47
*** nati_ueno has joined #openstack-ceilometer03:25
*** terriyu has quit IRC03:34
*** matsuhashi has quit IRC03:42
*** matsuhashi has joined #openstack-ceilometer03:43
*** ddieterly has quit IRC03:45
*** matsuhashi has quit IRC03:47
*** nosnos has quit IRC03:51
*** terriyu has joined #openstack-ceilometer04:00
*** nati_ueno has quit IRC04:08
*** nati_ueno has joined #openstack-ceilometer04:09
*** deepthi has joined #openstack-ceilometer04:12
*** nati_ueno has quit IRC04:29
*** matsuhashi has joined #openstack-ceilometer04:31
*** nosnos has joined #openstack-ceilometer04:37
*** terriyu has quit IRC04:41
boris-42ildikov hi04:47
*** ajc_ has joined #openstack-ceilometer04:59
*** Underyx|off has joined #openstack-ceilometer05:03
*** Underyx|off is now known as underyx05:03
*** underyx is now known as Guest1390905:03
*** ddieterly has joined #openstack-ceilometer05:04
*** fnaval has quit IRC05:05
*** ddieterly has quit IRC05:08
*** Guest13909 is now known as Underyx|off05:14
*** psharma has joined #openstack-ceilometer05:16
*** nati_ueno has joined #openstack-ceilometer05:17
*** rwsu has quit IRC05:24
*** nati_ueno has quit IRC05:26
*** ildikov has quit IRC05:27
*** zqfbot has quit IRC05:30
*** zqfbot has joined #openstack-ceilometer05:33
*** ddieterly has joined #openstack-ceilometer05:34
*** ddieterly has quit IRC05:38
*** yatin has joined #openstack-ceilometer05:44
*** ildikov has joined #openstack-ceilometer05:44
ildikovboris-42: hi06:05
*** inc0_ has joined #openstack-ceilometer06:21
*** _nadya_ has joined #openstack-ceilometer06:25
*** ddieterly has joined #openstack-ceilometer06:34
*** ddieterly has quit IRC06:39
*** psharma_ has joined #openstack-ceilometer06:42
*** nati_ueno has joined #openstack-ceilometer06:42
*** psharma__ has joined #openstack-ceilometer06:45
*** psharma has quit IRC06:45
*** psharma_ has quit IRC06:49
*** psharma_ has joined #openstack-ceilometer06:52
*** psharma__ has quit IRC06:55
*** _nadya_ has quit IRC07:01
boris-42ildikov hi there one more time=)07:01
*** Qiming has quit IRC07:02
*** Underyx|off is now known as underyx07:03
boris-42ildikov hm I finished spec for my patch
boris-42ildikov lol it's bigger then patch=)07:05
*** _cjones_ has joined #openstack-ceilometer07:09
ildikovboris-42: haha :) it can happen nowadays :)07:10
*** _cjones_ has quit IRC07:10
ildikovboris-42: I saw that you uploaded the spec, I will try to have a look at today07:11
*** _cjones_ has joined #openstack-ceilometer07:11
boris-42ildikov thank you07:11
ildikovboris-42: np!07:12
*** yatin has quit IRC07:13
*** Qiming has joined #openstack-ceilometer07:17
*** yatin has joined #openstack-ceilometer07:17
*** ifarkas has joined #openstack-ceilometer07:20
*** yatin has quit IRC07:23
*** ddieterly has joined #openstack-ceilometer07:34
*** Longgeek has joined #openstack-ceilometer07:37
*** Longgeek has quit IRC07:37
*** Longgeek has joined #openstack-ceilometer07:37
*** _nadya_ has joined #openstack-ceilometer07:38
*** ddieterly has quit IRC07:39
*** ddieterly has joined #openstack-ceilometer07:46
*** ddieterly has quit IRC07:50
*** matsuhashi has quit IRC07:57
*** matsuhashi has joined #openstack-ceilometer07:58
*** _cjones__ has joined #openstack-ceilometer08:00
*** IvanBerezovskiy has joined #openstack-ceilometer08:01
*** _cjones_ has quit IRC08:03
*** matsuhashi has quit IRC08:05
*** matsuhashi has joined #openstack-ceilometer08:10
*** matsuhashi has quit IRC08:17
*** matsuhashi has joined #openstack-ceilometer08:18
*** amarao has joined #openstack-ceilometer08:18
*** nijaba has joined #openstack-ceilometer08:21
*** nijaba has quit IRC08:22
*** nijaba has joined #openstack-ceilometer08:22
*** _cjones__ has quit IRC08:23
*** _cjones_ has joined #openstack-ceilometer08:23
*** yatin has joined #openstack-ceilometer08:24
*** isviridov|away is now known as isviridov08:26
*** matsuhashi has quit IRC08:31
*** Ju_ has quit IRC08:34
*** _cjones_ has quit IRC08:40
*** _cjones_ has joined #openstack-ceilometer08:41
*** _cjones_ has quit IRC08:45
*** eglynn has joined #openstack-ceilometer08:45
*** nosnos has quit IRC08:45
*** openstackgerrit has joined #openstack-ceilometer08:52
*** nijaba has quit IRC08:56
*** _nadya_ has quit IRC09:03
DinaBelovaeglynn, o/ Just read your msg to amalagon about "Options for Period-Spanning Statistics in Gnocchi" doc - what's this?09:12
DinaBelovaeglynn, sorry, I think I missed this somehow..09:12
eglynnDinaBelova: good morning!09:15
eglynnDinaBelova: ... that's just a write-up that Ana is working on09:15
DinaBelovaeglynn, good morning sir!09:15
eglynnDinaBelova ... of options for computing moving-averages, exponential smoothing and the likes, over gnocchi time-series09:15
DinaBelovaeglynn, oh, so that's not published yet?09:15
eglynnDinaBelova: (her OPW project is based on that idea)09:15
eglynnDinaBelova: ... no it's just a rough WIP deriving from an IRC conversation earlier in the week09:16
DinaBelovaeglynn, oh, ok!09:16
eglynnDinaBelova: ... but I can cerrainly ask her to share it more widely if you're interested?09:16
DinaBelovaeglynn, yes, please!09:16
DinaBelovaeglynn, really it's interesting topic09:16
eglynnDinaBelova: np! ... she'll prolly want to polish a bit more, but additional eyes would be good09:17
DinaBelovaeglynn, and I guess it might be really helpful09:17
DinaBelovaeglynn, thank you sir09:17
openstackgerritTatsuro Makita proposed a change to openstack/ceilometer: Add log output of "x-openstack-request-id" from nova
DinaBelovaeglynn, btw, strangely our openstackgerrit bot is failing really frequently - SergeyLukjanov had to restart it yesterday evening and this morning too :-\09:19
eglynnDinaBelova: yeah, is there a separate bot per project? ... or a single global one?09:20
eglynnDinaBelova: ... we've hardly been working it to death recently09:20
DinaBelovaeglynn, it's one shared bot09:21
DinaBelovabut sometimes it losts some of the channels09:21
DinaBelovaand ceilometer channel is lucky one :D09:21
eglynnDinaBelova: ... just outta interest, did you get OpenTSDB up & running?09:21
* eglynn is just casting around for realistic comparison points with gnocchi and infux ...09:21
* eglynn notes with interest ...09:22
eglynn"OpenTSDB stores data at the original resolution indefinitely, requesting data for a long time span can return millions of points."09:22
DinaBelovaeglynn, not myself09:22
eglynn... it's also made outta Java09:22
eglynn... ickkk!09:22
DinaBelovait's hbase based09:22
DinaBelovathat's why it has all pros and cons of hbase :)09:23
DinaBelovait's fast, scalable, good working with the time-series data (due to built-it tim fields, etc)09:23
DinaBelovabut it's not so easy as Mongo, for instance09:23
eglynn... the hbase-tax ;)09:24
DinaBelovahehe, indeed09:24
eglynnthe point in their docco about data retention was interesting09:24
eglynnpolicy seems to be ... keep everything09:24
eglynn... but also provide simple downsampling of query results09:25
eglynnalso their notion of aggregation is a little different to ours09:25
eglynnfor OpenTSDB, aggregation is a way of *combining* datapoints from multiple time-series09:26
eglynn(as opposed to rolling up data to a lower resultion *within* a single time-series)09:26
eglynnWTF is "resultion" eglynn?09:26
DinaBelovaeglynn, sorry, did not get your "WTF is "resultion" eglynn?"09:27
eglynnDinaBelova: just talking to myself, poking fun at my chronic inability to spell stuff ...09:28
DinaBelovaheh, sorry :D09:28
DinaBelovaas for the openstdb09:28
DinaBelovayes, they have other solutions that are chosen by pandas, for instance09:28
eglynn"other solutions" for metric storage?09:29
DinaBelovaI'm about the aggregation stuff that's built-in09:29
eglynna-ha, k09:29
DinaBelovanot pandas, probably, but current gnocchi stuff - wrong word combination, sorry09:30
DinaBelovabut still it has lots of advantages - speaking about large scale thing09:30
DinaBelovaand I found the python client for them09:31
eglynnDinaBelova: you've lost me there on the "current gnocchi stuff"?09:31
boris-42eglynn pls take a look at my spec =)09:31
*** nati_ueno has quit IRC09:31
eglynnboris-42: yep, it's on my list09:31
boris-42eglynn thank you09:31
*** safchain has joined #openstack-ceilometer09:32
DinaBelovaeglynn, nah, sorry, I just mean that current gnocchi == pandas + swift with our own solutions about aggregation, etc. if it'll be opentsdb backend, we'll need to check that all our 'vocabulary' is ok to be used in the opentsdb world09:32
DinaBelovaeglynn, but i guess it'll be ok :)09:33
eglynnDinaBelova: a-ha, yes, I see what you mean ... concern about possible "impedance mismatch" between aggregation models09:34
*** _nadya_ has joined #openstack-ceilometer09:34
DinaBelovaeglynn, yes - as we have chosen one of them, opentsdb uses other one09:34
DinaBelovabut anyway, we may extend what do we name 'aggregation'09:35
DinaBelovaand make this backend-specific, for instance09:35
DinaBelovaor find more common meaning for this inside our code09:35
DinaBelovaeglynn, btw -
eglynnDinaBelova: so IIUC, our concept of aggregation is more akin to the OpenTSDB concept of downsampling09:35
DinaBelovaeglynn, yes - like making dataset smaller09:36
*** nosnos has joined #openstack-ceilometer09:36
DinaBelovawell, not the dataset, but 'some amount of data is the average meaning'09:36
*** matsuhashi has joined #openstack-ceilometer09:36
DinaBelovatheir aggregation is like 'combination'09:37
DinaBelovasorry, dunno how to name this stuff better09:37
*** nijaba has joined #openstack-ceilometer09:37
*** nijaba has quit IRC09:37
*** nijaba has joined #openstack-ceilometer09:37
eglynntheir tagging notion is interesting too ...09:38
eglynne.g. a datapoint of form ... cpu_util time0 42% host=X instance=Y
*** Alice911 has quit IRC09:39
eglynnthe host, instance, tags then form different dimensions along which these data can be "aggregated"09:39
DinaBelovaeglynn, yes, and it's flexible enough imho09:39
*** Alice911 has joined #openstack-ceilometer09:40
eglynn... so I guess the interesting thing is we don't currently have a gnocchi concept analogous to tags09:41
DinaBelovaeglynn, yes, we just have one way of aggregation now09:41
DinaBelovaand no, well, ways to choose what and how to aggregate09:42
*** Alice911 has quit IRC09:42
DinaBelovaI guess that is future it might be some common interface - that will have different capabilities for different backends09:42
DinaBelovaeglynn, imho opentsdb is a better solution (comparing with the influxdb, for instance), as that's more stable and well-supported thing09:44
DinaBelovaas it's not the new one in this world09:45
eglynnDinaBelova: ... I've been thinking that we'll need to tackle some form of dimensionality in gnocchi anyway09:45
eglynnDinaBelova: ... so *something* around that will be needed I think as a "core capability"09:45
eglynnDinaBelova: ... re. OpenTSDB versus InfluxDB09:45
DinaBelovaeglynn, indeed - probably we can use 'tagging' naming here, btw09:46
DinaBelovait's a good one09:46
eglynnDinaBelova: ... we'll have one of the main influx contributors attending in Paris next week09:46
DinaBelovaeglynn, cool!09:46
eglynnDinaBelova: ... so we could get him to do a compare-and-contrast with OpenTSDB09:46
DinaBelovayes, it'll be simply great09:47
DinaBelovaeglynn, btw :)09:47
DinaBelovaI have goood news :)09:47
eglynnDinaBelova: ... I'm sure he encountered that question once or twice before09:47
DinaBelovaI got visa :)09:47
eglynnDinaBelova: cool, the 4am trip not wasted so :)09:47
DinaBelovaeglynn, heh, yes, comparing here will be anyway09:47
DinaBelovaeglynn, yes, I feel happy :D09:47
*** _nadya_ has quit IRC09:52
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Fix E251 violations and re-enable gating
*** Qiming has quit IRC09:53
SergeyLukjanoveglynn, btw trusty will be the default slave soon09:54
SergeyLukjanoveglynn, probably today :)09:54
eglynnSergeyLukjanov: yeah that great news!09:54
*** matsuhashi has quit IRC09:54
eglynnSergeyLukjanov: ... mtreinish & clarkb were talking about it last night at the project/release meeting09:55
SergeyLukjanoveglynn, yup09:55
*** matsuhashi has joined #openstack-ceilometer09:55
* eglynn looks forward to full gating against mongodb finally! :)09:55
SergeyLukjanoveglynn, I've been extremely sleepy at this meething, so, duplicating :)09:55
eglynnSergeyLukjanov: nah, it is major step forward, /me is a happy man! :)09:56
ildikovSergeyLukjanov: duplicating good news is always accepted :)09:56
*** yatin has quit IRC09:57
eglynnSergeyLukjanov: ... though of course the downside for me09:57
SergeyLukjanoveglynn, ildikov ;)09:57
eglynnSergeyLukjanov: ... is that it somewhat cuts the legs out from under one motivation for using Fedora20 more widely in the gate09:57
*** matsuhas_ has joined #openstack-ceilometer09:57
*** matsuhashi has quit IRC09:57
eglynnSergeyLukjanov: ... but hey, that's an argument for another day09:57
eglynnSergeyLukjanov: ... BTW presumably you'll be in Darmstadt at the qa/infra meetup?09:58
SergeyLukjanoveglynn, I'm don't think that fedora will be popular for gate due to the too many releases :)09:58
eglynnSergeyLukjanov: (... there will also be a couple Red Hatters there beating the f20/centos7 drum)09:58
SergeyLukjanoveglynn, I'm not planning going to the Darmstadt right now09:58
eglynnSergeyLukjanov: ... a-ha, k09:58
SergeyLukjanoveglynn, but I'll be on your sprint with Dina :)09:59
*** underyx is now known as Underyx|off09:59
eglynnSergeyLukjanov: excellent! :)09:59
SergeyLukjanoveglynn, are you going to Darmstadt?10:01
eglynnSergeyLukjanov: probably not, but there are a couple other Red Hatters going10:02
eglynnSergeyLukjanov: ... unfortunately that's not likely to include ianw though10:02
eglynnSergeyLukjanov: ... logistics of travelling from .au are too prolly awkward for him10:03
eglynnSergeyLukjanov: ... but Attila and Pavel will be there and on hand to address some of the f20-related concerns10:03
eglynnSergeyLukjanov: (... e.g. the starvation issue in the nodepool grant allocation aglo)10:04
*** _nadya_ has joined #openstack-ceilometer10:18
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Fix E265 violations and re-enable gating
SergeyLukjanoveglynn, cool10:19
DinaBelovaeglynn, hehe, I *love* this chain :D
DinaBelovathat's the top commit there :)10:21
DinaBelovabetter code - better project :)10:21
*** amarao has quit IRC10:22
eglynn... and more OCD-prone coders ;)10:22
DinaBelovaeglynn :D10:23
*** _nadya_ has quit IRC10:27
*** _nadya_ has joined #openstack-ceilometer10:28
*** matsuhas_ has quit IRC10:28
*** matsuhashi has joined #openstack-ceilometer10:28
*** Underyx|off is now known as underyx10:30
*** ifarkas has quit IRC10:32
*** matsuhashi has quit IRC10:37
*** matsuhashi has joined #openstack-ceilometer10:37
*** yatin has joined #openstack-ceilometer10:50
eglynnSergeyLukjanov: nice :)10:59
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Fix F402 violations and re-enable gating
DinaBelovaeglynn, btw, are there any webinar records?11:28
DinaBelovaI had no opportunity to attend, but I really would like to go through it ^)11:28
*** yatin has quit IRC11:29
eglynnDinaBelova: slides ...
eglynnDinaBelova: audio will be on youtube in a couple days apparently11:29
DinaBelovaeglynn, thanks a lot!11:29
eglynnDinaBelova: ... on this channel
DinaBelovaeglynn, thank you!11:41
*** yatin has joined #openstack-ceilometer11:44
*** erecio has joined #openstack-ceilometer11:44
*** dmsimard_away is now known as dmsimard11:49
openstackgerritLadislav Smola proposed a change to openstack/ceilometer: Enhancing test data script for UI testing
*** amarao has joined #openstack-ceilometer11:55
DinaBelovalsmola__, o/ won't be good for you in the future? :)11:56
lsmola__DinaBelova: I started to poke about the other script
DinaBelovahehe, I see this :)11:58
DinaBelovaok, I'm glad the current one meets all your needs :)11:58
lsmola__DinaBelova: I will comment the sample generator, seems like those two should have similar interface11:58
DinaBelovalsmola__, I guess it'll be nice - please feel free to do this :)12:00
lsmola__DinaBelova: yeah, I will just make some minor enhancement and it should be enough for shellscript that will set complete set of meters for UI testing12:00
lsmola__DinaBelova: btw. for us we will probably want to see how the UI looks like with year of data12:01
lsmola__DinaBelova: which will be quicker if we put it directly into DB12:01
*** terriyu has joined #openstack-ceilometer12:01
DinaBelovalsmola__, what do you mean here?12:06
DinaBelovasorry, I may not understand you correctly12:06
DinaBelovawhat kind of data do you want to see presented in the db?12:06
lsmola__DinaBelova: like the other script is sending it through messaging system, so setting large number of samples will eb slower12:07
DinaBelovaa-ha, so smth like emulating huge amount of data already presented in the db?12:07
lsmola__DinaBelova: basically we want to see how it will look like in Production12:08
lsmola__DinaBelova: yeah12:08
DinaBelovalsmola__, btw, are you interested in the performance testing results?12:08
DinaBelovafor the ceilo?12:08
lsmola__DinaBelova: yeah12:08
DinaBelovacurrently we have the results for the icehouse cut12:08
lsmola__DinaBelova: nice, it would be nice to know the limits12:09
DinaBelovaand ityaptin is working on testing the current code and future vectors12:09
DinaBelovalsmola__, one moment12:09
DinaBelovalsmola__, we have tested the api and collectors working12:09
DinaBelovalsmola__, here it is -
DinaBelovathat was done for the icehouse code, currently MySQL, for instance, will look much better :)12:10
lsmola__DinaBelova: nice, thank you, this will be handy12:11
DinaBelovalsmola__, np12:11
*** erecio has quit IRC12:11
*** erecio has joined #openstack-ceilometer12:12
DinaBelovaone of the ceilo bottlenecks (for the hbase and mongo at least) is the queue - we even could not create load huge enough that will meet DB max throughput, the problem was in the slow queue presented12:13
*** erecio has quit IRC12:13
DinaBelovalsmola__, ^^ - that's why we're working on the oslo.messaging improvement as well12:13
*** erecio has joined #openstack-ceilometer12:14
*** alexpilotti has joined #openstack-ceilometer12:15
*** jdob has joined #openstack-ceilometer12:16
lsmola__DinaBelova: right, I can see that can be kind of bottleneck12:18
*** erecio has quit IRC12:20
*** erecio has joined #openstack-ceilometer12:21
*** promulo has quit IRC12:23
*** ajc_ has quit IRC12:23
*** amalagon has joined #openstack-ceilometer12:32
*** deepthi has quit IRC12:36
*** yatin has left #openstack-ceilometer12:39
*** Qiming has joined #openstack-ceilometer12:54
*** prad_ has joined #openstack-ceilometer12:56
*** _nadya_ has quit IRC12:57
*** matsuhashi has quit IRC12:59
*** matsuhashi has joined #openstack-ceilometer12:59
*** julim has joined #openstack-ceilometer13:03
*** matsuhashi has quit IRC13:04
*** safchain has quit IRC13:06
*** safchain has joined #openstack-ceilometer13:06
*** thomasem has joined #openstack-ceilometer13:10
*** matsuhashi has joined #openstack-ceilometer13:10
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Fix H307 violations and re-enable gating
*** fnaval has joined #openstack-ceilometer13:21
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Fix F402 violations and re-enable gating
*** matsuhashi has quit IRC13:33
*** matsuhashi has joined #openstack-ceilometer13:33
*** underyx is now known as Underyx|off13:34
*** matsuhashi has quit IRC13:38
jd__I'm getting crazy with those statistics stuff13:42
*** psharma_ has quit IRC13:43
DinaBelovajd__, what's up?13:43
*** nosnos has quit IRC13:43
jd__I'm tired, what about you DinaBelova?13:43
DinaBelovajd__, hehe :D13:43
openstackgerritRen Qiaowei proposed a change to openstack/ceilometer-specs: Spec for xenapi support
DinaBelovame too, but I have good news :)13:43
*** terriyu has quit IRC13:43
DinaBelovaI'll be in Paris :)13:44
jd__ah awesome13:44
*** matsuhashi has joined #openstack-ceilometer13:44
DinaBelovajd__, try not to become too exhausted :( that's bad :(13:44
*** matsuhashi has quit IRC13:44
jd__yeah I'll sleep at some point :)13:45
*** matsuhashi has joined #openstack-ceilometer13:45
DinaBelovajd__, good luck!13:47
*** matsuhashi has quit IRC13:49
amalagonjd__: if you need an extra pair of hands to do something, let me know!13:54
*** promulo has joined #openstack-ceilometer13:54
*** Ruetobas has quit IRC13:54
jd__amalagon: thanks :) I think I'm waiting for eglynn to lighten me on how we should handle various things13:54
jd__cf the mail thread13:54
* amalagon reading the thread13:55
*** nati_ueno has joined #openstack-ceilometer13:55
*** fnaval has quit IRC13:57
*** Ruetobas has joined #openstack-ceilometer14:01
eglynnjd__: sorry for being such a pain, labouring that pairwise-cumulative point to death14:04
jd__no problem eglynn14:05
jd__we need to have that discussion to progress14:05
*** Ruetobas has quit IRC14:05
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Fix H307 violations and re-enable gating
eglynnjd__: ... will reply to your latest on that thread shortly14:06
*** ildikov has quit IRC14:06
jd__cool thanks14:06
jd__I'm writing tooz documentation in the meantime14:06
jd__you know, to chill out14:06
eglynnjd__: LOL :)14:06
*** ddieterly has joined #openstack-ceilometer14:09
*** Ruetobas has joined #openstack-ceilometer14:12
*** nati_ueno has quit IRC14:16
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Fix H307 violations and re-enable gating
*** sileht has quit IRC14:20
*** sileht has joined #openstack-ceilometer14:23
eglynnjd__: just sent ... apols for the delay, /me on a boring internal call14:24
*** amalagon_ has joined #openstack-ceilometer14:24
*** IvanBerezovskiy has left #openstack-ceilometer14:28
* DinaBelova still thinking what will happen with swift (that actually has different endpoints in replicas for the some object) - I mean with replicas... at one time we have no possibility to guarantee it'll be exactly one file on every replica - it'll be consistent finally, but not at every moment14:28
DinaBelovajd__, probably I'm wrong here ^^, I'm not really sure...14:29
*** ildikov has joined #openstack-ceilometer14:30
eglynnDinaBelova: eventually consistency would be less of a potential issue if the only operations where ...14:30
eglynn(1) write a new datapoint14:31
eglynn(2) bulk expiry of old datapoints14:31
eglynn(i.e. no roll-up/downsampling in situ)14:31
DinaBelovaeglynn, yes, I know this, this idea just came to my mind14:32
DinaBelovaas datapoints may be added time by time - from the ceilo itself, they'll be collected periodically14:33
DinaBelovajust meters14:33
DinaBelovafor instance14:33
*** gordc has joined #openstack-ceilometer14:33
DinaBelovaso if it'll be request - give all new datapoints14:34
DinaBelovasome new ones might not be presented14:34
DinaBelovaas we dunno what swift server will actually perform the request14:35
*** fnaval has joined #openstack-ceilometer14:38
*** _nadya_ has joined #openstack-ceilometer14:38
eglynnDinaBelova: ... so first let's simplify and assume *no downsampling*14:39
*** fnaval has quit IRC14:39
*** fnaval has joined #openstack-ceilometer14:39
eglynnDinaBelova: ... datapoints {p1, p2, ..., p_i} are submitted and are stored in an eventually consistent way14:40
eglynnDinaBelova: ... possible issue 1: a query returns only {p1, p2, ..., p_(i-1)}14:40
eglynnDinaBelova: ... ^^^is that the way swift eventual consistency actually works?14:41
eglynnDinaBelova: ... in any case, I'm not sure that's a big issue ... the caller just sees slightly laggy, but not *wrong* data14:41
DinaBelovaeglynn, I'm not the specialist here (will investigate this moment), but as far as I remember, swift first of all updates one server, and then shares the data. So yes, that might look like this. And yes, that' not the biggest problem here14:42
DinaBelovathat'll be the opportunity to get not updated data14:42
eglynnDinaBelova: ... possible issue 2: a further datapoint {p_(i+1)} arrives and is folded into {p1, p2, ..., p_(i-1)} instead of {p1, p2, ..., p_i}14:42
eglynnDinaBelova: ... issue #2 would be the real concern I think14:43
DinaBelovaeglynn, sorry, I need to test this somehow...14:43
DinaBelovaI'm not sure what kind of situation we'll have here14:43
DinaBelova#1 I guess we'll meet for sure, but #2 - not so..14:44
eglynnDinaBelova: we have an expert in the house! :) ...
DinaBelovaeglynn, I guess I14:44
DinaBelovawill ask jd__when he'll stop dying because of cumulative aggregations :)14:45
*** rwsu has joined #openstack-ceilometer14:49
*** terriyu has joined #openstack-ceilometer14:52
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Fix F402 violations and re-enable gating
*** terriyu has quit IRC14:54
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Fix F402 violations and re-enable gating
*** fnaval has quit IRC15:12
*** fnaval has joined #openstack-ceilometer15:13
*** idegtiarov has quit IRC15:15
*** fnaval has quit IRC15:17
*** terriyu has joined #openstack-ceilometer15:22
*** terriyu has quit IRC15:29
*** terriyu has joined #openstack-ceilometer15:29
*** terriyu has quit IRC15:29
*** fnaval has joined #openstack-ceilometer15:37
*** _nadya_ has quit IRC15:40
*** _nadya_ has joined #openstack-ceilometer15:58
*** Qiming has quit IRC16:00
*** Ruetobas has quit IRC16:01
*** Ruetobas has joined #openstack-ceilometer16:03
*** Ruetobas has quit IRC16:08
*** Ruetobas has joined #openstack-ceilometer16:14
*** _cjones_ has joined #openstack-ceilometer16:18
*** _cjones_ has quit IRC16:19
*** _cjones_ has joined #openstack-ceilometer16:20
*** packet has joined #openstack-ceilometer16:24
*** safchain has quit IRC16:56
*** amarao has quit IRC16:59
*** eglynn has quit IRC17:07
*** jaypipes has joined #openstack-ceilometer17:17
*** rbowen has joined #openstack-ceilometer17:39
*** jaypipes has quit IRC17:44
*** jaypipes has joined #openstack-ceilometer17:45
*** eglynn has joined #openstack-ceilometer17:46
*** fnaval has quit IRC17:52
*** nati_ueno has joined #openstack-ceilometer18:00
*** Qiming has joined #openstack-ceilometer18:01
*** Qiming has quit IRC18:09
*** ddieterly has quit IRC18:10
*** inc0_ has quit IRC18:18
*** ddieterly has joined #openstack-ceilometer18:24
*** ddieterly has quit IRC18:28
*** ddieterly has joined #openstack-ceilometer18:29
*** isviridov is now known as isviridov|away18:32
*** ddieterly has quit IRC18:33
*** ddieterly has joined #openstack-ceilometer18:34
*** ddieterly has quit IRC18:38
*** ddieterly has joined #openstack-ceilometer18:38
*** ddieterly has quit IRC18:43
*** ddieterly has joined #openstack-ceilometer18:43
*** ddieterly has quit IRC18:47
*** ddieterly has joined #openstack-ceilometer18:48
*** ddieterly has quit IRC18:50
*** ddieterly has joined #openstack-ceilometer18:50
*** amalagon_ has quit IRC18:52
*** _nadya_ has quit IRC18:54
*** ddieterly has quit IRC18:55
*** aviau has quit IRC18:55
*** ddieterly has joined #openstack-ceilometer18:55
*** ddieterly has quit IRC18:58
*** ddieterl_ has joined #openstack-ceilometer18:58
*** Fabio_ has joined #openstack-ceilometer19:00
*** openstackgerrit has quit IRC19:04
*** Longgeek has quit IRC19:04
*** openstackgerrit has joined #openstack-ceilometer19:05
*** yjiang5 is now known as yjiang5_away19:08
*** Longgeek has joined #openstack-ceilometer19:08
*** aviau has joined #openstack-ceilometer19:10
*** amalagon_ has joined #openstack-ceilometer19:14
*** fnaval has joined #openstack-ceilometer19:16
*** liusheng has quit IRC19:20
*** liusheng has joined #openstack-ceilometer19:20
*** _nadya_ has joined #openstack-ceilometer19:24
*** alexpilotti has quit IRC19:32
*** amalagon_ has left #openstack-ceilometer19:44
*** _cjones_ has quit IRC19:44
*** _cjones_ has joined #openstack-ceilometer19:45
*** ddieterl_ has quit IRC19:46
*** erecio has quit IRC19:52
*** erecio has joined #openstack-ceilometer19:57
*** openstackgerrit has quit IRC20:05
*** Longgeek has quit IRC20:05
*** openstackgerrit has joined #openstack-ceilometer20:05
*** ddieterly has joined #openstack-ceilometer20:12
*** amarao has joined #openstack-ceilometer20:17
*** Underyx|off is now known as underyx20:19
*** underyx is now known as Underyx|off20:26
*** ivan has joined #openstack-ceilometer20:27
*** Underyx|off is now known as underyx20:29
*** promulo has quit IRC20:50
boris-42eglynn ildikov guys did you have a chance to take a look at my spec?20:51
*** yjiang5_away is now known as yjiang520:52
*** erecio has quit IRC20:56
openstackgerritPhil Neal proposed a change to openstack/ceilometer-specs: PaaS Usage Event Collection
eglynnboris-42: ... I have now21:03
boris-42eglynn thanks21:03
ildikovboris-42: I had a quick look, not a heavy weight docco for sure ;)21:05
* eglynn scurries off to watch France grind out a nil-nil with Ecuador21:10
*** jdob has quit IRC21:10
eglynn... considers it prep for the mid-cycle in Paris next week ;)21:10
boris-42eglynn replied to your comment21:12
ildikoveglynn: ... yeap, better to be prepared :)21:12
*** promulo has joined #openstack-ceilometer21:15
*** amarao has quit IRC21:16
*** julim has quit IRC21:16
*** julim has joined #openstack-ceilometer21:20
*** julim has quit IRC21:24
openstackgerritBoris Pavlovic proposed a change to openstack/ceilometer-specs: Add spec for osprofiler notification plugin
boris-42eglynn ^ addressed your issues21:26
*** gordc has quit IRC21:34
*** jaypipes has quit IRC21:34
*** rbowen has quit IRC21:47
*** _nadya_ has quit IRC21:52
*** thomasem has quit IRC21:52
*** ivan has quit IRC21:54
ildikovboris-42: eglynn: posted a question on that spec21:56
boris-42ildikov ?21:56
*** ivan has joined #openstack-ceilometer21:57
boris-42ildikov i added my answers=)21:57
boris-42ildikov hmmm21:57
boris-42ildikov how to do that?)21:57
ildikovboris-42: it should be visible now21:58
boris-42ildikov new table?21:58
ildikovboris-42: yeap, I was thinking about that21:59
boris-42ildikov hm and how to do that wihtotu copy paste of existing notification plugins ?)21:59
boris-42ildikov I mean implementation will require a lot of code22:00
boris-421) changes in API and client22:00
ildikovboris-42: I think it will need some additional code, I do not really see an extra simple solution for that right now22:00
boris-422) adding one more framework for notifications plugins22:01
boris-423) changes in DB22:01
boris-424) a lot of tests & as well we will need functional tests (that will be not so simple to implement)22:01
boris-42not sure that it's better then just to add notification plugin22:02
*** yjiang5 is now known as yjiang5_away22:02
*** yjiang5_away has quit IRC22:02
ildikovI know, I'm just a bit concerned about to mix this data completely22:02
boris-42ildikov hm but event types are different22:03
boris-42ildikov so you can query data quite simple22:03
ildikovwe have a ttl mechanism for samples for instance, which may should be different in case of profiling data22:04
boris-42ildikov ttl?22:04
*** nikhil___ is now known as nikhil___|afk22:04
boris-42ildikov you mean temporary samples?22:04
ildikovwell, it depends on what you call temporary :)22:05
boris-42ildikov that will be deleted after some period of time (automatically)22:06
ildikovthen yes22:06
boris-42ildikov not sure that I would like to use it for profiling stuff22:06
boris-42ildikov I would prefer to have API method to delete them22:07
boris-42ildikov but not do this automatically22:07
ildikovsure, so it is one point, where profiling data should be handled differently22:08
boris-42ildikov probably we should just be able to delete samples by query?22:09
boris-42ildikov I am afraid actually about writing something custom for profiler22:09
ildikovit's another sensitive question related to billing22:09
boris-42eglynn ^ what do you think about all this ?22:10
boris-42ildikov heh I like KISS approach =)22:10
ildikov... so maybe we could provide this possibility to the admin user, but still questionable22:11
boris-42ildikov actually it will be very useful22:12
boris-42ildikov for operators22:12
boris-42ildikov for example you would like to check how your cloud works22:12
ildikovheh, sure, we can keep it simple from coding PoV, but it's not the only thing :)22:12
boris-42ildikov under load22:12
boris-42ildikov so you run rally create billions of samples22:12
boris-42ildikov pre production testing22:13
boris-42ildikov and now you would like to cleanup all of them22:13
*** ddieterly has quit IRC22:14
*** dmsimard is now known as dmsimard_away22:15
ildikovI know it would be useful in general, I mean profiling22:15
ildikovthe only thing is that in my opinion it would be better to separate the data and data handling from metering22:16
ildikov... so I raised this question now, better earlier than later IMHO22:16
boris-42ildikov sure22:16
*** yjiang5_1 has joined #openstack-ceilometer22:18
ildikovso I like profiler, I'm just simply the member of the Ceilo guardians team in this discussion, so I try to act like that :)22:18
*** _nadya_ has joined #openstack-ceilometer22:19
boris-42ildikov okay going to sleep now=) see you tomorrow22:19
ildikovgood idea22:21
ildikovI guess eglynn practices how to celebrate France at the moment :)22:21
ildikov... so let's continue this tomorrow22:21
*** ddieterly has joined #openstack-ceilometer22:24
*** ddieterly has quit IRC22:26
*** ddieterly has joined #openstack-ceilometer22:26
*** openstackgerrit has quit IRC22:35
*** openstackgerrit has joined #openstack-ceilometer22:35
*** nsaje has quit IRC23:01
*** nsaje has joined #openstack-ceilometer23:05
*** ddieterly has quit IRC23:21
*** ddieterly has joined #openstack-ceilometer23:22
*** ddieterly has quit IRC23:23
*** fnaval has quit IRC23:44
*** r-daneel has joined #openstack-ceilometer23:48
*** r-daneel has quit IRC23:52

Generated by 2.14.0 by Marius Gedminas - find it at!