 Yeah, so hello everybody Welcome to the cloud kitty product update. My name is Luca pesce. I'm the ptls for the stein release I've been contributing since me tarka and I've been a core since three or four releases. I believe You are a lot of people Just for the record who here is using or has used cloud kitty in the past. Could you please raise your hands? Nobody, okay Perfect That's actually a good thing so What is cloud kitty cloud kitty is a rating and charge back component for open stock It takes your gnocchi or monaska or promity use metrics and applies a rating rule to them Which gets committed to a storage backend and are available for you to retrieve be grouped and get Lot of nicer information It is not a billing solution, which means you can't apply taxes with cloud kitty You can convert euros to dollars what we do is basically take your metrics collect them and the way you want a play rating rules to them and Generate CSV reports, which can be injected in your own homegrown billing solution or your whatever software you are using We have a very basic horizon integration and more the more comes from the New storage backend, which will be introduced later We are compatible with influx DB now, which means you are you can build graphana dashboards for your customers or Internal teams whoever you charge your cloud usage to so the rocky release has seen a few new features More important one are the internal refactoring and Cleanup so a lot of things have been removed. So legacy code has been removed Like the Cilometer collector so you can't use a Cilometer API anymore, which is deprecated since Ocarta. I believe The storage has been refactored, but we will talk about this later on and A new client is available, which is a Python library should be easier to use And is integrated with OpenStack client. So yes, if you see OpenStack rating, it's us And most of the most important part cloud kitty can be used in standalone mode now Which means you don't need an open-stack cloud and a keystone endpoint to use cloud kitty for example a customer of our company has Used gnocchi in standalone mode with VMware and a collector which pushed the VMware metrics into gnocchi And cloud kitty was used to process these VMware metrics, but it could be anything So any custom metrics which you push into gnocchi or Prometheus or Madasca, but I don't know if people do this can be rated with Cloud kitty actually What's upcoming for OpenStack stein a v2 storage interface? actually this interface is available in Rocky, but you shouldn't use it. It's unstable. It's only for development purposes it's meant to Tackle the same issues the telemetry project encountered a while ago Which means we have a huge amount of data to store and well it causes problems as always So this backend aims at be more flexible on that side and to do nice things like a filtering grouping Saying what are my top 10 most expensive volumes for the given period on that domain that kind of queries Which means we need a v2 API to expose this new internal format This API will be based on flask rather than a pecan which will make contributions a lot easier because everybody knows flask We will try to focus on the community This had been this had been a it has been a subject For the stein cycle and will be important in the future We would like to get more feedback and to make the community grow So we would like to gain some contributors to have an external point of view to have several use cases I don't know a huge scale or a very specific rating or well, we would like to have a bigger diversity and contributors from all around the world and They have always also been improvements for the standard on mode The from it use collector. Well Has been has been published We know how the scope notion in the past you had to group your your data on a tenant basis Now that you can use close it in the standard on mode. You can group your data on anything domain user Data center ID for a huge people anything And there is a gnocchi fetcher. So that's something very specific basically In the past you had to apply the rating rule To the tenants you wanted to rate Now you can do a dynamic tenant discovery through gnocchi because in most clouds people actually want to rate everything and Every new tenant which is created. So this can be done now when you use a gnocchi beyond stein The v2 API will be extended We want to extend it slowly so endpoint for endpoint and we would like people to make their contributions to Well, tell us about their needs and to add their endpoints which match their needs or demands and Once this v2 API is considered complete and stable in a few releases It will become the default and yes, and we will deprecate the v1 API We would like to create a new rating model The funny thing is when you talk to different customers Each one has his ID of what is an Intuitive API to push rating rules and everybody sees this thing differently. So Implementing a new rating model isn't really a technical challenge, but it is a human challenge Because it's very hard to make something which is easy to use for everybody and which everybody understands easily It will continue the focus on the community We recently reintroduced the IRC meetings We will push the usage of specs, which means every new feature will go through a through a spec in the cloud kitty specs repo and Well, we'd like the changes to be really announced community wide We'll also make a heavy use of the mailing list. Well heavy use than right now And we have put some focus on the community goals Which means that we are Python 3 compatible And we provide well cylinder review, but we should provide upgrade checks I don't know if you have heard about that specific community goal First time every project will provide upgrade check tool which will allow to verify if your Deployment in version n is ready for an upgrade to version n plus one or if you should change things If you'd like to give feedback or to contribute Please remember that any contribution is welcome, which means you can code you can review But you can also tell us you can open bugs. You can tell us. Okay. I have this specific need You can work on the documentation Anything is actually a good contribution We've got an onboarding session in the room M1 which is in this direction on this level At 5 10 p.m. We are available on IRC on three nodes We have a group on storyboard. We use a storyboard for reviews, etc and this is my personal email and Well, Claude Kitty at objective Libre is an alias to all the to all the Claude Kitty Developers at my company. Do we have any questions? I hope to do there was a microphone here either use it or 10 year question. I will repeat it for the record So the question is what is the big difference to Sly clubs, which is also doing open-stack building metrics? Claude Kitty is an old project. Actually it started in Kilo, I believe and it's a big 10 project. So it's unofficial open-stack project and We actually want to be fully open source So it's not to make money for our company It's really to contribute to the open-stack project because that's how the company makes its living and yes, basically we're completely open source and You can easily add your own rating model If there are contributors and open-stack different developers here, you probably have heard of Steve door Any component rating module collector a storage back-end is loaded with Steve door Which allows you to write your own storage back-end collector rating module API, whatever This is some other question Yes So the question was can you use only prometheus for data collection? Yes, you can use a cloud kitty outside of open-stack and standalone mode collect are you rating information from prometheus and push it to The storage back-end which is for the moment either v1 sql alchemy or v2 influx db Which gives you all the nice v2 API features and a graph on integration. No other questions Yes So The question was if the monastic integration is production ready We tested it in production at the small scale So about a thousand instances and 30 hypervisors, so it's running It was back ported to me taca in an HPE context. So HP helium It is but if you want to deploy it We are on IRC and tell us how it goes for you and we would be glad to Maintain this monaska integration the monaska guys have been really help from an IRC when we coded this so Any further question? Well, thank you very much, and I hope to see you at the project onboarding