| *** ducttape_ has joined #openstack-monasca | 00:31 | |
| *** bobh_ has joined #openstack-monasca | 00:48 | |
| *** bobh_ has quit IRC | 00:53 | |
| *** efoley has joined #openstack-monasca | 01:11 | |
| *** ducttape_ has quit IRC | 01:32 | |
| *** bmotz has quit IRC | 01:42 | |
| *** bmotz has joined #openstack-monasca | 01:42 | |
| *** bobh_ has joined #openstack-monasca | 01:49 | |
| *** bobh_ has quit IRC | 01:52 | |
| *** bobh_ has joined #openstack-monasca | 01:52 | |
| *** efoley has quit IRC | 01:55 | |
| *** ljxiash has joined #openstack-monasca | 01:55 | |
| *** bobh_ has quit IRC | 02:01 | |
| *** bobh_ has joined #openstack-monasca | 02:20 | |
| *** ukaynar_ has joined #openstack-monasca | 02:26 | |
| *** ukaynar_ has quit IRC | 02:30 | |
| *** ukaynar_ has joined #openstack-monasca | 02:30 | |
| *** vishwanathj has joined #openstack-monasca | 02:32 | |
| *** vishwanathj has quit IRC | 02:32 | |
| *** bobh_ has quit IRC | 03:02 | |
| *** bobh_ has joined #openstack-monasca | 03:25 | |
| *** bobh_ has quit IRC | 03:29 | |
| *** bobh_ has joined #openstack-monasca | 04:25 | |
| *** bobh_ has quit IRC | 04:31 | |
| *** ukaynar_ has quit IRC | 04:32 | |
| *** ljxiash has quit IRC | 05:29 | |
| *** pcaruana has joined #openstack-monasca | 05:39 | |
| *** bobh_ has joined #openstack-monasca | 06:03 | |
| *** bobh_ has quit IRC | 06:07 | |
| *** witek has joined #openstack-monasca | 06:48 | |
| *** ljxiash has joined #openstack-monasca | 08:48 | |
| *** rbak has quit IRC | 08:59 | |
| *** pcaruana has quit IRC | 09:00 | |
| *** nadya has joined #openstack-monasca | 09:02 | |
| *** bobh_ has joined #openstack-monasca | 09:09 | |
| *** kei_yama has quit IRC | 09:13 | |
| *** kei_yama has joined #openstack-monasca | 09:13 | |
| *** bobh_ has quit IRC | 09:14 | |
| *** hosanai has quit IRC | 09:41 | |
| openstackgerrit | Koji Nakazono proposed openstack/monasca-log-api: Add devstack plugin installer https://review.openstack.org/300396 | 09:54 |
|---|---|---|
| openstackgerrit | Tomasz Trębski proposed openstack/monasca-common: Periodicity for metrics https://review.openstack.org/292753 | 10:07 |
| *** nadya has quit IRC | 11:09 | |
| *** bobh_ has joined #openstack-monasca | 11:10 | |
| *** bobh_ has quit IRC | 11:14 | |
| *** nadya has joined #openstack-monasca | 11:47 | |
| *** vaquerog has joined #openstack-monasca | 11:56 | |
| *** kei_yama has quit IRC | 12:06 | |
| *** kei_yama has joined #openstack-monasca | 12:11 | |
| *** Gen has joined #openstack-monasca | 12:11 | |
| *** Gen has quit IRC | 12:11 | |
| *** Gen has joined #openstack-monasca | 12:12 | |
| *** kse has quit IRC | 12:16 | |
| *** Gen has quit IRC | 12:34 | |
| *** kei_yama has quit IRC | 12:37 | |
| *** Gen has joined #openstack-monasca | 12:58 | |
| *** ljxiash has quit IRC | 13:36 | |
| *** ddieterly has joined #openstack-monasca | 13:49 | |
| *** ljxiash has joined #openstack-monasca | 13:49 | |
| *** bobh_ has joined #openstack-monasca | 13:55 | |
| *** bobh_ has quit IRC | 14:00 | |
| *** hosanai has joined #openstack-monasca | 14:05 | |
| *** nadya has quit IRC | 14:07 | |
| *** rbak has joined #openstack-monasca | 14:13 | |
| *** ddieterly is now known as ddieterly[away] | 14:21 | |
| *** craigbr has joined #openstack-monasca | 14:23 | |
| *** ddieterly[away] is now known as ddieterly | 14:23 | |
| *** rhochmuth has joined #openstack-monasca | 14:24 | |
| *** Fdaisuke_ has joined #openstack-monasca | 14:24 | |
| *** tongli has joined #openstack-monasca | 14:28 | |
| *** vaquerog_ has joined #openstack-monasca | 14:28 | |
| *** suksant has joined #openstack-monasca | 14:30 | |
| rhochmuth | #topic monasca-analytics | 14:30 |
| *** vaquerog has quit IRC | 14:31 | |
| hosanai | hello | 14:32 |
| rhochmuth | hi hisashi | 14:33 |
| openstackgerrit | Ryan Bak proposed openstack/monasca-agent: Rework agent checks to use multiprocessing library https://review.openstack.org/301355 | 14:33 |
| rhochmuth | luis should be here in a minute | 14:33 |
| hosanai | rhochmuth: hi. OK we will wait him. | 14:33 |
| Fdaisuke_ | Hi, | 14:33 |
| rhochmuth | hi | 14:33 |
| Fdaisuke_ | I work with hosanai at same company. | 14:34 |
| hosanai | rhochmuth: Fdaisuke_ is my coleague i think are you daisuke? | 14:34 |
| rhochmuth | cool | 14:34 |
| vaquerog_ | Hi gentlemen, this is Luis | 14:34 |
| *** david______ has joined #openstack-monasca | 14:34 | |
| hosanai | vaquerog_: hi! | 14:34 |
| *** david______ has quit IRC | 14:34 | |
| vaquerog_ | Suksant is in the channel too, from the Labs team in the UK (and David will be joining as well) | 14:35 |
| rhochmuth | hi luis, are you the only one from bristol? | 14:35 |
| rhochmuth | cool, typed to quick | 14:35 |
| rhochmuth | hi suksant | 14:35 |
| suksant | Hi | 14:35 |
| rhochmuth | looks like we have a good group | 14:35 |
| rhochmuth | should we wait for anyone else? | 14:35 |
| hosanai | maybe david? | 14:35 |
| rhochmuth | so is this a good meeting time? | 14:36 |
| vaquerog_ | david is here in the same room, so we can probably go ahead. He'll be on in 2 secs | 14:36 |
| *** david_subiros_pe has joined #openstack-monasca | 14:36 | |
| david_subiros_pe | hi | 14:36 |
| rhochmuth | hi david | 14:36 |
| rhochmuth | i think that is everyone | 14:36 |
| rhochmuth | welcome to the wayback machine | 14:37 |
| vaquerog_ | :) | 14:37 |
| hosanai | i'm so sorry for the last meeting | 14:37 |
| *** craigbr has quit IRC | 14:37 | |
| rhochmuth | hosanai, that was my fault | 14:37 |
| rhochmuth | so, where should we begin | 14:38 |
| rhochmuth | i'm still in the process of submitting your repo for opensstack | 14:38 |
| rhochmuth | actually, i had to do another one | 14:38 |
| vaquerog_ | :'( sorry to hear that | 14:38 |
| rhochmuth | but, now that monasca-transform is done, i'll submit monasca-anlytics today | 14:38 |
| rhochmuth | it want' worth submitting both at the same time, because io would jsut make the same mistakes twice | 14:39 |
| rhochmuth | but, now i'm all clear | 14:39 |
| rhochmuth | i think | 14:39 |
| rhochmuth | so, i'll get that done today | 14:39 |
| hosanai | cool! | 14:39 |
| rhochmuth | it will still proabbly take a week or two to get approved | 14:39 |
| vaquerog_ | Excellent! | 14:39 |
| rhochmuth | because the repo is a part of an official project, the governance committee needs to approve | 14:40 |
| rhochmuth | i think | 14:40 |
| vaquerog_ | We are almost done and waiting for our legal guys to approve the release | 14:40 |
| rhochmuth | but, i don't expect any issues | 14:40 |
| hosanai | rhochmuth: til approval, can we see the repo? | 14:40 |
| rhochmuth | so, the final approval should hopefully be just a formality | 14:40 |
| *** ducttape_ has joined #openstack-monasca | 14:40 | |
| rhochmuth | no, they create it | 14:40 |
| vaquerog_ | I will, since we are mostly building on open source tools | 14:40 |
| rhochmuth | i believe | 14:41 |
| hosanai | i see. | 14:41 |
| rhochmuth | so, in the mean time, i think we would have to create a public github repo somewhere | 14:41 |
| vaquerog_ | we can sync up and I can give you an overview before approval | 14:41 |
| rhochmuth | we could do that | 14:41 |
| rhochmuth | takes like 5 minutes | 14:41 |
| vaquerog_ | so that we can start discussing next steps and roadmap as a team | 14:41 |
| rhochmuth | so luis, that sounds good | 14:41 |
| hosanai | sounds good! | 14:42 |
| rhochmuth | if you want to give an overview, that woudl be great | 14:42 |
| vaquerog_ | at the moment we have a thin orchestration layer on top of Spark | 14:42 |
| vaquerog_ | that simply controls how data and results flow bewteen different algorithms | 14:42 |
| vaquerog_ | we plan to expand that to Flink | 14:42 |
| vaquerog_ | they say that it kicks spark's bottom out in most performance tests | 14:42 |
| vaquerog_ | plus, our experience is that spark is very knobby and there are many levers to be adjusted | 14:43 |
| rhochmuth | hmmm, i hadn't seen that | 14:43 |
| vaquerog_ | any thoughts on that? | 14:43 |
| rhochmuth | i had a very brief look at flink | 14:43 |
| rhochmuth | the monasca-transform engine that is currently in progress is based on spark right no | 14:44 |
| vaquerog_ | I think these slides gave more details http://es.slideshare.net/sbaltagi/flink-vs-spark | 14:44 |
| rhochmuth | it hasn't been formally reviewed yet | 14:44 |
| *** craigbr has joined #openstack-monasca | 14:44 | |
| rhochmuth | ok, i'll look at off-line | 14:44 |
| rhochmuth | as you know, the threshold engine is storm | 14:44 |
| hosanai | thanks for the info i will read it later. | 14:45 |
| rhochmuth | so, you've got storm, spark, samaz and now flink to look at | 14:45 |
| vaquerog_ | well lets assume we just build on spark for the time being | 14:45 |
| *** rbrndt has joined #openstack-monasca | 14:45 | |
| vaquerog_ | on top of that we have this thin orchestration layer that simply deals with data handling and sending the right algorithm | 14:45 |
| vaquerog_ | on top of the thin orchestration there is all the intelligence | 14:46 |
| vaquerog_ | we have a lightweight dataflow language that specifies how data, algorithms and results are put together | 14:46 |
| vaquerog_ | for instance | 14:46 |
| vaquerog_ | you can use 4 different algorithms to infer causality on alertsa | 14:46 |
| vaquerog_ | you can run thos ein parallel | 14:46 |
| vaquerog_ | but you want the output to be a combination of those | 14:47 |
| vaquerog_ | meaning that if "A" is caused by "B" in all algorithms | 14:47 |
| hosanai | i would like to know more about dataflow language. what is it? | 14:47 |
| vaquerog_ | ten, the aggregate output should have a high certainty that is true | 14:47 |
| vaquerog_ | at the moment is just a json description | 14:47 |
| vaquerog_ | I think we can share that with you even without legal approval | 14:48 |
| vaquerog_ | again | 14:48 |
| vaquerog_ | we can change oit, refine it | 14:48 |
| vaquerog_ | or go to some sort of DSL that we can use more programmatically | 14:48 |
| vaquerog_ | any thoughts on that? what would be your preference? | 14:48 |
| rhochmuth | i think json is fine for now | 14:49 |
| rhochmuth | and over-time can look into something more advanved | 14:49 |
| vaquerog_ | ok, that sounds good | 14:50 |
| rhochmuth | hosanai, any preference | 14:50 |
| hosanai | so far, json is fine for us. but i would like to know from monasca point of view. | 14:51 |
| hosanai | what is best for technically? | 14:51 |
| hosanai | s/what/which/ | 14:51 |
| vaquerog_ | Not sure I fully understand the question, but let me try to ive an answer | 14:52 |
| rhochmuth | so, my understanding is that currently there is a json config file to define the pipeline | 14:52 |
| vaquerog_ | the algorithmic flow would be consumed solely by monasca-analytics | 14:52 |
| rhochmuth | right | 14:52 |
| vaquerog_ | yes, you define the data processing pipeline there and the thin orchestration relies that to spark/flink | 14:53 |
| rhochmuth | so, when monasca-analytics starts-up, it reads the pipeline definition | 14:53 |
| vaquerog_ | from the Monasca standpoint, I presume it makes little difference | 14:53 |
| rhochmuth | correct | 14:53 |
| vaquerog_ | you can specify kafka as a source of the dataflow | 14:53 |
| *** bobh_ has joined #openstack-monasca | 14:53 | |
| hosanai | sorry, i understand my mis-understanding | 14:53 |
| vaquerog_ | also as a sink so that anyone can dump/consume data into the flow | 14:53 |
| vaquerog_ | we will provide a couple of examples | 14:54 |
| vaquerog_ | one on how to do anomaly detection (an example of flow that can be run on monasca-analytics) | 14:54 |
| vaquerog_ | and another one on alert aggregation | 14:54 |
| vaquerog_ | for instance, the result of an alert aggregation flow can be dumped on vitrage for further manipulation | 14:55 |
| hosanai | what is vitrage? | 14:55 |
| rhochmuth | vitrage is a new project in openstack for root cause canalysis | 14:55 |
| rhochmuth | it uses topologicial information to do rca | 14:56 |
| vaquerog_ | https://github.com/openstack/vitrage/ | 14:56 |
| rhochmuth | and there will be two presentaitons in Austin | 14:56 |
| hosanai | rhochmuth: thanks! i will check it. | 14:56 |
| rhochmuth | unfortunately, one overlaps with monasca design summit related discussions | 14:56 |
| vaquerog_ | it is just an example of consumer of one of the 9very many) possible algorithmic flows | 14:56 |
| vaquerog_ | sorry, Roland, could you elaborate on that a bit? | 14:57 |
| rhochmuth | the overlap? | 14:57 |
| vaquerog_ | yes, please | 14:57 |
| rhochmuth | now that monasca is an official openstack project there will be design summit sessions in Austin | 14:58 |
| *** craigbr has quit IRC | 14:58 | |
| rhochmuth | monstly, they are all wednesday morning, and one wednesday agernoon | 14:58 |
| vaquerog_ | ok, I see | 14:58 |
| vaquerog_ | thanks | 14:58 |
| rhochmuth | so, you had mentioned to me several algorithms | 14:58 |
| rhochmuth | via email | 14:58 |
| rhochmuth | i was wondering what the state of your current development in that area was | 14:59 |
| vaquerog_ | yes, at the moment we will provide two examples on alert causality inference and anomaly detection | 14:59 |
| vaquerog_ | we plan to make them a bit more complex | 14:59 |
| vaquerog_ | we believe the community would be an excellent place to: | 14:59 |
| vaquerog_ | 1) share data (or data samplees) | 15:00 |
| vaquerog_ | 2) share algorithms | 15:00 |
| rhochmuth | any other references, so hosanai and Fdaisuke_ can get started | 15:00 |
| rhochmuth | or doe we will have to wait for the osrb approval | 15:00 |
| vaquerog_ | the second one is interesting since we can (over time) create a library of algorithmics based on best practices or success stories for the benefit of the community | 15:00 |
| rhochmuth | yes, i agree, | 15:01 |
| vaquerog_ | we can point you to the libraries and papers we used to build those | 15:01 |
| rhochmuth | that would be great | 15:01 |
| rhochmuth | i believe you sent me some linke, which i reviewed | 15:01 |
| hosanai | great! | 15:01 |
| rhochmuth | it looked good | 15:01 |
| rhochmuth | but, would be good to post again here or via email | 15:01 |
| *** ducttape_ has quit IRC | 15:01 | |
| david_subiros_pe | this is the anomaly detection algorithm that we are using: http://scikit-learn.org/stable/modules/generated/sklearn.svm.OneClassSVM.html | 15:02 |
| Fdaisuke_ | So, cool | 15:02 |
| *** bmotz_ has joined #openstack-monasca | 15:02 | |
| vaquerog_ | as you can imagine all of them are basd on unsupervised learning techniques | 15:02 |
| *** craigbr has joined #openstack-monasca | 15:02 | |
| *** ducttape_ has joined #openstack-monasca | 15:02 | |
| vaquerog_ | part of the roadmap we'd like to discuss is techniques for labelling data and beenfiting from semi-supervised/supervised techniques | 15:03 |
| vaquerog_ | reconginsing cats in youtube is easy | 15:03 |
| hosanai | i see unsupervised one is suitable for our use cases. | 15:03 |
| vaquerog_ | my 2 year old daughter can do that without millions of samples to train her (Google always go for the easy problems :P ) | 15:03 |
| vaquerog_ | openstack is more complex and there is not an abundance of experts with time to collaborate | 15:04 |
| vaquerog_ | that is why unsupervised is highly relevant here | 15:04 |
| vaquerog_ | and semi-supervised can be part of the roadmap, if you think that is a good idea and could be useful | 15:04 |
| vaquerog_ | any thoughts? | 15:04 |
| hosanai | do you have a plan to use supervised tech in future? | 15:05 |
| rhochmuth | i think unsupervised is where to start | 15:05 |
| rhochmuth | we dont' have large datasets available that are classified for anomalies | 15:05 |
| suksant | for causality, we try LiNGAM (we implemented it based on scikit-learn) | 15:05 |
| *** bmotz has quit IRC | 15:05 | |
| rhochmuth | or alarm clustering | 15:05 |
| hosanai | rhochmuth: we don't have ORACLE (great teacher) :-) | 15:06 |
| *** dschroeder has joined #openstack-monasca | 15:06 | |
| suksant | https://sites.google.com/site/sshimizu06/lingam | 15:06 |
| vaquerog_ | well, we were thinking of using things ike label propagation to walk that huge gap (for instance, if you have a graph of connected samples, some of them anomalous, you could do some label propagation) | 15:06 |
| vaquerog_ | and throw that to an admin to guide the algorithmic flow | 15:06 |
| vaquerog_ | What suksant just added is the algorithm we will be giving as an example of causality | 15:07 |
| vaquerog_ | what datasets could we use today? | 15:07 |
| rhochmuth | so getting data is the problem | 15:08 |
| rhochmuth | i can get data for our internal clouds from the last time i did this | 15:08 |
| hosanai | sorry, causality means kind of correlation analytics. i coun't follow it. | 15:08 |
| rhochmuth | but, that is all the data i have access to these days | 15:08 |
| vaquerog_ | We could also start analysing the openstack build and test environment | 15:08 |
| vaquerog_ | I guess some tests pass and when you deploy in production you find odd stuff going on | 15:09 |
| rhochmuth | i'm wondering if twc could provide data from thier environment | 15:09 |
| vaquerog_ | thworing some analytics at that dataset may save many hours | 15:09 |
| rhochmuth | the openstack build/test environment doesnt' use monasca | 15:09 |
| rhochmuth | not sure what historical data they have avaialble | 15:10 |
| rhochmuth | for metrics and alarms | 15:10 |
| hosanai | rhochmuth: does monasca use in OIC? | 15:10 |
| rhochmuth | hosanai: i thought you also had algorithms that fujitsi had developed? | 15:11 |
| rhochmuth | sorry, i don't know what oic is | 15:11 |
| hosanai | rhochmuth: wait a minute i will check the correct name. | 15:11 |
| rhochmuth | luis, how soon do you think osrb approval will occur | 15:12 |
| hosanai | OpenStack Innovasion center | 15:12 |
| rhochmuth | thanks hosanai. | 15:12 |
| rhochmuth | i'm not aware of anything going on in there | 15:13 |
| suksant | we started the process few weeks ago. i will follow up with the responsible person. | 15:13 |
| rhochmuth | never even heard of it before | 15:13 |
| hosanai | in lab we have some algorithms but i think we will not use them in monasca | 15:13 |
| rhochmuth | thanks suksant | 15:13 |
| rhochmuth | so, how do we get coordinated more, if can't release code and no repo yet? | 15:14 |
| rhochmuth | is there anything we can do in the mean time | 15:14 |
| vaquerog_ | We'll send the flow spec to get feedback | 15:14 |
| rhochmuth | any code that can be released? It doesnt' sound like it | 15:14 |
| rhochmuth | ok | 15:14 |
| rhochmuth | can i add that to the monasca wiki? | 15:15 |
| vaquerog_ | and I can show that presentations (an expanded version of what we have just discussed) using lync or something similar | 15:15 |
| *** ljxiash has quit IRC | 15:15 | |
| rhochmuth | ok | 15:15 |
| *** craigbr has quit IRC | 15:15 | |
| vaquerog_ | we expect to be able to release in a few weeks | 15:15 |
| rhochmuth | thanks | 15:15 |
| rhochmuth | so, i'll have the repo ready to go | 15:16 |
| rhochmuth | by then | 15:16 |
| vaquerog_ | we'll chase the attorneys (we usually run away from them :) | 15:16 |
| rhochmuth | lol | 15:16 |
| vaquerog_ | Hisashi, your feedback is VERY important for us | 15:16 |
| hosanai | Fdaisuke_ and I would like to play (test) with the code first to understand it. | 15:16 |
| vaquerog_ | it would be great if you coul dshare your plans, concerns and what you'd like us to work together on | 15:17 |
| vaquerog_ | that makes sense, but we dould ditch our code and start afresh if it doesn't meet your requirements | 15:17 |
| vaquerog_ | I think it would be good to understand those a little bit mroe | 15:17 |
| *** openstackgerrit has quit IRC | 15:18 | |
| rhochmuth | hosanai and Fdaisuke_: do you have any code that is avaialble from you end for review? | 15:18 |
| rhochmuth | or algorithms and descriptions? | 15:18 |
| *** openstackgerrit has joined #openstack-monasca | 15:18 | |
| hosanai | rhochmuth: we don't have it. our current plan is understand luis's code then check differences | 15:19 |
| rhochmuth | thx | 15:19 |
| vaquerog_ | not having code is fine for us | 15:20 |
| vaquerog_ | but understanding desired features, requirements, and use cases would be great | 15:20 |
| rhochmuth | so, how about i create a landing page for monasca-analytics in the monasca wiki | 15:21 |
| hosanai | vaquerog_: ok, i will prepare them based on our POC's info | 15:21 |
| rhochmuth | and we can start to identiy some of those use case | 15:21 |
| rhochmuth | and we can start to post information and collaboration more on this | 15:22 |
| vaquerog_ | Great! Thanks!! | 15:22 |
| rhochmuth | i'll get somethign in-place today | 15:22 |
| hosanai | rhochmuth: it's a good idea ! | 15:22 |
| hosanai | oh, too quick :-) | 15:22 |
| rhochmuth | i was wondering what led to your selection of the svm for outlier detection? | 15:23 |
| vaquerog_ | Excellent! We can then put together a list of requirements and share more broadly to define our roadmap | 15:23 |
| rhochmuth | so, i've got two action items. 1 repo, 2 wiki | 15:24 |
| vaquerog_ | well, we tried a few | 15:24 |
| vaquerog_ | there is not too many open source options in unsupervised | 15:24 |
| vaquerog_ | well, not ones that are computationally efficient | 15:24 |
| rhochmuth | hmmm, did you consider non-parametric statistics | 15:25 |
| vaquerog_ | yes, but didn't have time to explore it further. Most of our efforts were derailed into building the thin orchestration and the interpreter for the flow | 15:25 |
| rhochmuth | i see | 15:26 |
| david_subiros_pe | we tried this one as well: https://unsupervisedlearning.wordpress.com/2014/08/04/topological-anomaly-detection/ but got better results with SVM one class. | 15:26 |
| rhochmuth | thx david | 15:26 |
| rhochmuth | i'll review that one too | 15:26 |
| david_subiros_pe | Maybe we can try more options, and depending on the data source we can use different algorithms | 15:26 |
| rhochmuth | what types of things did you use svm for? | 15:27 |
| vaquerog_ | kernel density anomaly detection is one clear candidate to be added | 15:27 |
| rhochmuth | there were some interesting papers by twitter last year on this topic | 15:27 |
| rhochmuth | i'll dig them up | 15:27 |
| hosanai | could you share links for it? | 15:27 |
| rhochmuth | or jsut googel twitter breakout detection and anomaly detection | 15:27 |
| david_subiros_pe | We used SVM one class to find anomalies in firmware/hardware level | 15:27 |
| rhochmuth | i'll have to find again | 15:27 |
| vaquerog_ | ok, we'll do | 15:27 |
| rhochmuth | i was working on portign the breakout detection to python | 15:28 |
| rhochmuth | that is funny, my ull request is still there at, https://github.com/twitter/BreakoutDetection/pulls | 15:28 |
| rhochmuth | here is the github repo, https://github.com/twitter/BreakoutDetection | 15:29 |
| rhochmuth | the underlying code is all C/C++ | 15:29 |
| rhochmuth | they have an R binding | 15:29 |
| rhochmuth | i tried to add Python | 15:29 |
| rhochmuth | but then broke the R | 15:29 |
| rhochmuth | the python port works | 15:29 |
| vaquerog_ | :) We love breaking things | 15:29 |
| hosanai | lol | 15:29 |
| *** craigbr has joined #openstack-monasca | 15:29 | |
| rhochmuth | there is a bloq or paper on that method | 15:30 |
| rhochmuth | and then there was a follow-up for anomaly detection | 15:30 |
| suksant | This one I suppose https://github.com/twitter/AnomalyDetection | 15:30 |
| rhochmuth | the stuff that i had previousely done using Komagorov-Smirnov two sample tests | 15:31 |
| rhochmuth | suksant: correct, that is the one | 15:31 |
| rhochmuth | it looks like they aren't actively working on it | 15:31 |
| rhochmuth | but it should be good to review and use | 15:32 |
| rhochmuth | so, i have to run to another meeting | 15:32 |
| rhochmuth | should we schedule another meeting for next week | 15:33 |
| vaquerog_ | ok, | 15:33 |
| rhochmuth | or two weeks? | 15:33 |
| vaquerog_ | we'll send you the json flow descriptor | 15:33 |
| rhochmuth | actually, in two weeks i'll be at openstack summit | 15:33 |
| rhochmuth | so, probably next week | 15:33 |
| vaquerog_ | and we can present the slides I mentioned above | 15:33 |
| vaquerog_ | ok | 15:33 |
| vaquerog_ | next week then | 15:33 |
| rhochmuth | sounds good | 15:33 |
| hosanai | rhochmuth: ok. thanks & have a nice day! | 15:34 |
| rhochmuth | i'll schedule it in outlook | 15:34 |
| rhochmuth | so you next week same time | 15:34 |
| vaquerog_ | cheers | 15:34 |
| vaquerog_ | bye | 15:34 |
| Fdaisuke_ | ok, Thank you in advance. | 15:34 |
| rhochmuth | thanks eveyrone! | 15:34 |
| david_subiros_pe | bye, thanks! | 15:34 |
| suksant | thank you. bye. | 15:34 |
| *** vaquerog_ has quit IRC | 15:34 | |
| *** david_subiros_pe has quit IRC | 15:34 | |
| *** suksant has quit IRC | 15:34 | |
| openstackgerrit | ZhiQiang Fan proposed openstack/python-monascaclient: Implement monasca metric-name-list https://review.openstack.org/303683 | 15:34 |
| *** rbrndt__ has joined #openstack-monasca | 15:38 | |
| *** rbrndt has quit IRC | 15:39 | |
| *** _ducttape_ has joined #openstack-monasca | 15:45 | |
| *** ducttape_ has quit IRC | 15:47 | |
| *** hosanai has quit IRC | 15:48 | |
| *** ddieterly is now known as ddieterly[away] | 15:48 | |
| *** ddieterly[away] is now known as ddieterly | 15:51 | |
| openstackgerrit | Merged openstack/monasca-api: Dimension Filtering in Vertica https://review.openstack.org/292654 | 16:04 |
| openstackgerrit | Bradley Klein proposed openstack/monasca-api: Replace measurements table join with subquery https://review.openstack.org/304208 | 16:10 |
| *** openstackstatus has joined #openstack-monasca | 16:11 | |
| *** ChanServ sets mode: +v openstackstatus | 16:11 | |
| *** craigbr has joined #openstack-monasca | 16:11 | |
| -openstackstatus- NOTICE: Reminder, Gerrit will be offline from 20:00 to 21:00 UTC for a server replacement http://lists.openstack.org/pipermail/openstack-dev/2016-April/091274.html | 16:19 | |
| openstackgerrit | Bradley Klein proposed openstack/monasca-api: Replace measurements table join with subquery that fits in memory https://review.openstack.org/304208 | 16:19 |
| openstackgerrit | Bradley Klein proposed openstack/monasca-api: Replace measurements table join with subquery that fits in memory https://review.openstack.org/304208 | 16:53 |
| *** Fdaisuke_ has quit IRC | 16:54 | |
| *** nadya has joined #openstack-monasca | 17:18 | |
| *** ddieterly is now known as ddieterly[away] | 17:31 | |
| *** craigbr has quit IRC | 17:40 | |
| *** ddieterly[away] is now known as ddieterly | 17:53 | |
| *** craigbr has joined #openstack-monasca | 17:56 | |
| *** ddieterly is now known as ddieterly[away] | 18:16 | |
| *** craigbr has quit IRC | 18:22 | |
| *** craigbr has joined #openstack-monasca | 18:26 | |
| openstackgerrit | David Schroeder proposed openstack/monasca-agent: A couple small fixes to the postfix plugin https://review.openstack.org/304276 | 18:30 |
| *** nadya has quit IRC | 18:34 | |
| *** vishwanathj has joined #openstack-monasca | 18:37 | |
| *** vishwana_ has joined #openstack-monasca | 18:39 | |
| *** vishwanathj has quit IRC | 18:41 | |
| *** vishwana_ is now known as vishwanathj | 18:43 | |
| *** ddieterly[away] is now known as ddieterly | 18:44 | |
| *** rbrndt has joined #openstack-monasca | 18:55 | |
| *** rbrndt__ has quit IRC | 18:55 | |
| *** craigbr_ has joined #openstack-monasca | 19:04 | |
| *** craigbr has quit IRC | 19:04 | |
| *** craigbr_ is now known as craigbr | 19:04 | |
| *** ddieterly is now known as ddieterly[away] | 19:27 | |
| *** nadya has joined #openstack-monasca | 19:30 | |
| *** craigbr has quit IRC | 19:46 | |
| *** ddieterly[away] is now known as ddieterly | 19:50 | |
| -openstackstatus- NOTICE: Gerrit will be offline from 20:00 to 21:00 UTC (starting 10 minutes from now) for a server replacement http://lists.openstack.org/pipermail/openstack-dev/2016-April/091274.html | 19:50 | |
| *** ddieterly is now known as ddieterly[away] | 19:52 | |
| *** craigbr has joined #openstack-monasca | 19:52 | |
| *** ddieterly[away] is now known as ddieterly | 19:55 | |
| *** nadya has quit IRC | 20:00 | |
| *** openstackgerrit has quit IRC | 20:05 | |
| -openstackstatus- NOTICE: Gerrit is offline until 21:00 UTC for a server replacement http://lists.openstack.org/pipermail/openstack-dev/2016-April/091274.html | 20:06 | |
| *** ChanServ changes topic to "Gerrit is offline until 21:00 UTC for a server replacement http://lists.openstack.org/pipermail/openstack-dev/2016-April/091274.html" | 20:06 | |
| *** openstackgerrit has joined #openstack-monasca | 20:28 | |
| *** ddieterly is now known as ddieterly[away] | 20:35 | |
| *** dschroeder has quit IRC | 20:36 | |
| *** openstackgerrit has quit IRC | 20:41 | |
| *** ddieterly[away] is now known as ddieterly | 20:42 | |
| *** openstackgerrit has joined #openstack-monasca | 20:55 | |
| *** dschroeder has joined #openstack-monasca | 21:00 | |
| *** ChanServ changes topic to "OpenStack Monitoring as a Service | https://wiki.openstack.org/wiki/Monasca" | 21:06 | |
| -openstackstatus- NOTICE: Gerrit move maintenance completed successfully; note that DNS has been updated to new IP addresses as indicated in http://lists.openstack.org/pipermail/openstack-dev/2016-April/091274.html | 21:06 | |
| openstackgerrit | Bradley Klein proposed openstack/monasca-api: Check gate/tempest with master https://review.openstack.org/304319 | 21:37 |
| openstackgerrit | Bradley Klein proposed openstack/monasca-api: Replace measurements table join with subquery that fits in memory https://review.openstack.org/304208 | 21:38 |
| *** craigbr has quit IRC | 21:45 | |
| *** ddieterly is now known as ddieterly[away] | 21:48 | |
| *** bobh_ has quit IRC | 21:55 | |
| *** stackedsax has quit IRC | 22:15 | |
| *** _ducttape_ has quit IRC | 22:19 | |
| *** ddieterly[away] is now known as ddieterly | 22:37 | |
| *** ddieterly is now known as ddieterly[away] | 22:55 | |
| *** ddieterly[away] is now known as ddieterly | 22:55 | |
| *** rbrndt has quit IRC | 22:59 | |
| *** rbak has quit IRC | 23:04 | |
| *** ddieterly is now known as ddieterly[away] | 23:05 | |
| *** ddieterly[away] has quit IRC | 23:06 | |
| *** tongli_ has joined #openstack-monasca | 23:22 | |
| *** tongli_ has quit IRC | 23:22 | |
| *** tongli_ has joined #openstack-monasca | 23:22 | |
| *** bobh_ has joined #openstack-monasca | 23:23 | |
| *** kse has joined #openstack-monasca | 23:24 | |
| *** tongli has quit IRC | 23:25 | |
| *** tongli_ has quit IRC | 23:27 | |
| *** kei_yama has joined #openstack-monasca | 23:29 | |
| *** rhochmuth has quit IRC | 23:37 | |
| *** tongli has joined #openstack-monasca | 23:47 | |
| *** hosanai has joined #openstack-monasca | 23:49 | |
| *** tongli has quit IRC | 23:51 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!