 So we were talking a bit about what you guys want to start today. Anything in particular? I do have a list. The order in which we proceed is a bit open. One thing I think would be top of mind is the Monoribo updates. I got an update from Joe on that. He said by then we should have a demo for the Monoribo, which is awesome. The Monoribo was done as part of our RFC process, and I thought this was a really great outcome from the Detroit 2022 Skiddinger Summit. So, I don't know, please, do you want to talk to any of that? The idea of the Monoribo, our motivations? I'm happy to as well. It's only in RFC, but my goal is to really just make sure that the Monoribo is real. And as I said, Joe, we were shooting forward, trying to get it ready before the stop. What we've just done by the end of the week is the Monoribo. There have been some specific challenges, but we're addressing a team of some things to do with some builder. There's a key that's still being in there, and so I have a follow-up term about some interesting questions and concerns, but I think it should be great. We use it now internally, so that's the RFC process. There's going to be questions about how we migrate from the existing to new, and it's up to do that, and that's really what the demo is about, showing we are in that conversation, how we make the journey that we're making this and everything. And then, if it's possible, it gives us the tool to move forward. I think it's going to be expression, free quality, so in our daily lives, internally, we're making forward to social growth. Yeah, I think having it happen, open source is going to be kind of the kick in the boat we need to go. So I'm going to move forward to where I'm happy to work with General Horvacher. So I'm going to understand all of the policy that isn't made well for him. And just for some context, for those who don't know, Spinnaker is made up of a bunch of different microservices. So currently, there's 10 or 11 microservices, and then there's a shared library called Cork. Any update that happens to Cork is propagated to each of the other resources. And then there's some of those other repositories like Orca or Fiat, that when they have updates, they in turn have more requests to other repos to update their libraries. This is a very complex, let's say, and by having an order with them, you'll have one request that would change the common library, as well as update all of those other microservices to make sure they're in line enough to date. This is going to really help with testing and making sure we're reducing breaking changes. I've been there myself where you have a previous, an older release that you need to make a change to the common library, and that change is just a lot of work. David's done some great work in automating the backporting changes for Cork to take that whole load off our plates, but just monitoring the work is really going to just help with building and release and reviewing changes as well. So we'll be able to see, hey, given this new feature, what are all the changes and all the services that are needed? That moves us to the next topic then, which is the automated testing topic. Right now, I mean, each repo has its own automated test, but they only go as far as they go. You know, we build Docker containers for Docker images, I should say, in each repo, but nothing ever, none of the automated processes ever actually spin up those containers. Spin up those images into containers and make sure they work, much less spinning them all up and make sure they work together. So that's a gap that it sounds like we're trying to fill, there are some UI testing efforts and some more backend-focused testing efforts. And I guess that was the next topic, is to see where people are with that and to try to make sure we're aligned and to coordinate. I think the goal is to have action that happens to pull requests and spins up the just built image and tries it out. Some things are going to be easier with the monorepo. I kind of hesitate to say that we're waiting for the monorepo because I don't really want to slow anything down. I think we just got some testing stuff. We'd love to have that help validate the monorepo. So, like I said, all, you know, full steam ahead of the fall, but you probably need to align. What do they need to spin up these Docker images? You know, you sort of quickly get it to like installation, how do you install spinning for conversations? I mean, it can inspire a lot of control, but it can also be super helpful to make this happen. On that topic, we have been working to kind of promote the test plan. I was going to say that this is the test that we've done. We're able to do that. We are going to open source our pipeline about leadership as well. That's probably the effort. We, yeah, I am just trying to make sure that I can go on and read some of all the things that have happened. The plan that's in place is good. I'm not just sharing that. But even with that, maybe you'll be able to do the same. We haven't actually made our public end. It's a recent final stages. The decision of the thinking is that, you know, in the 90s, you should be able to take off and get out of actions, bring up the instance of a spinnaker and put some tests into it. And then, you know, and they apply it. It's a cool test model. And then scaling. I've been building this in Java, so that was pretty much what I should do. It's pulling people to be further with it when they are in the action. Great. So there will be more probabilities than that? That's good. So you should be able to do this if you want to know where it comes from. Take this off. There aren't any 90s in the sun, there aren't any 90s right now. They're used to being in the old system. What we should get rid of, do you figure is going to host them? And so we have them on our board. You figure them out and you know it's not going to work. Because that's something you need to figure out. But you're going to be doing the test, right? So if you want to make it do it, thank you a lot. Right. So that's the place that's doing, that's doing the cat-hop. We're actually doing the resist. That's what happens. Okay. We want to scan it. Can you guys get it all up? Hi. We're having a platform signal. It's pretty formal. Okay. Yeah. We were just talking about some improvements we're making to a repo structure and a test. How can you, and we're all actually on the TLC, or something. Can you tell us, what's your interaction in the center for these days? Okay. Hi guys. So I joined Adobe about 10 years ago. Now we're getting a manager by an SRV team. And we are trying to create new pipelines for very old location that we have. The application right now is called Conex. It's kind of a platform. It's in little zoom. However, we're trying to build the next CT for this. And we are considering spin-off a lot. So in fact, we are joining PLC right now and it brought through questions from our team. And we're trying to understand if it can actually fit in your product, if it can automate. It's mainly actually a new workshop, even. We're moving to latest on a new platform. But it's going to take years to get there. So I want to try to understand more how spin-off is involving. We're not there. Basically, it's really complicated sometimes to finance. So that's my participation. I'm sorry. It's all in us now. It already has a cable to use in spin-off. It does have one. And I've heard from one of our engineers that they're moving away. Apparently because now they developed a platform that's called EPRS and is filled with ARP integrations. So there is that. I don't want to turn up too much for it. Because it's which is all we do. There's a lot of level commentators. So pipeline is complicated. So yeah, so we do a lot of Jenkins automation. But we think in a long run that would be very complicated to orchestrate a large pipeline using Jenkins. So we need this kind of orchestration on top of it. It sounds better. It's phase one now. They're applying and trying to create new environments to it. That's way better. First step is to bring the pipeline. What's that? And that's where we are right now. On the stage. So does it work well for you guys? Is that the future? How does it look in the near future now? For us it works great. But my quality is generalized. We have a lot of options. We have a lot of developers. And for use cases it's all for. It works great. The interesting thing about this landscape is for Kubernetes specific there are a lot of choices. But employees in everything manage some things across multiple different domains. And the hardware or some things you can use to make it go all the way. But I need to explain to everyone that you've got all these images. So that's what we're talking about. Yeah. What's the alternative that's going to go all the way now? To the player? Yeah. Yeah. Certainly to AWS Spinnaker is a good answer. Yeah. I need to be able to conceptually be able to do things for you guys. That's not a problem. Can I? Yeah. It can be hard to find resources. Spinnaker Slack and anybody can join and anybody can search for questions. Sometimes hard to get questions answered. It depends on what part of the life cycle you're at. If you're really the first time user of Spinnaker how do I get started? I can imagine it's a little daunting. Certainly when you get to expert level and you still run into issues you can ask much more specific questions. We probably have relationships with people that have been easier to get answers but longer in between. I kind of install a little sort of working but it's not exactly working. I mean I think that's one of the challenges we have with the community. Keeping the lights on and keep shipping personal software. I know for me I already invest a lot of time to start to invest more. So I kind of resist helping out on Slack sometimes. I don't know it's a challenge. In this one one of the key notes is how do you help the community and how do you get people to do something and all requests are not the only way to contribute for sure. Helping out on Slack. Helping out. It's great. You have issues on Slack. Those are the two places. I think it's still pretty It's like a challenge. Yeah, I think part of that too is we have this barrier to entry which is a really high bar entry which we have a lot of news that GSM is trying to explore that bar entry. There are some efforts going on with the customers and it's a little bit of a set-up shrinker and I think doing those things you know how do you guide some of that how do you help the community it's definitely going to be a process it's not going to be something that actually happens overnight it's going to have a lot of tech that and that's also the other thing here we are at the roadmap there's a lot of tech that which will also help you know more contributions because you know if you're using the Slayer's things you're going to want to actually be great. So as those things happen that we can keep momentum at least even in this group I think people see that improvement because it will make things easier for you to contribute it will make things easier to use that's what I'm talking about here if you have all the features you want to understand it they help you too but if we decrease the tech that it will be a lot of our entry so that when people have the knowledge and you'll see more of the response it comes very quickly so that's the workflow that we're trying to get to now all of those are still contributing things there's minus there's a few hours for the account handling through the dynamics you know everybody has some efforts if they are working on the contributing with these guys up here yeah and so that's right because it's more engaging there's things that you can use and it didn't make sense but the nature of the fact that we're going to do everything and also we'll keep you around for something going, going off of that too I think we were talking about integration and testing right before this and I think by having that in place it will make contributors more confident because it's pretty daunting when you make a change you have to look at other people's to identify and having good tests in place and making that transparent so that members of the community can see what happened, why did this test fail or I'm having this issue now can I write a test to exercise that to make sure we don't have that regression in the future so that type of participation has been really key in the community because everyone has a very, very good reason to enter and just that's talking with each other people use it in very, very different ways but it is flexible enough to do that and by having the different types of tests or workflows as part of that integration testing suite it will be great you just reminded me of something one of the things I mean like people do when they submit a code changer like hey this is broken I want to fix it or maybe I just keep saying this and I'm the annoying guy it's like well where's the test write a test that fails and automate a test that fails and then you fix it and then the test starts to pass test driven developments people won't even talk about any more because it's like the way people work but what Cameron is pointing out is that sometimes like some little detail like job level, super bits and bytes code test is not the test that people care about what people care about is like this pipeline needs to fail and if people, lots of people sort of know about pipelines probably then like intricacies of some you know crazy JVM detail and so to be able to submit a pipeline that fails is also another way to open things up to contributions one other thing I want to say though like I don't know how to decide to use Spinnaker or not to use Spinnaker one of the knocks I think on Spinnaker over the years has been that it's expensive to operate and I don't know it's definitely getting less expensive I'm going to give a talk tomorrow but it has some big numbers about how much less expensive it's getting but I work with Salesforce and we have pretty massive scale and we spend a lot of money on Spinnaker and people want us to spend a lot less money on Spinnaker so we I mean I have a whole team my team and cousin teams and sibling teams and you know hundreds as maybe slightly overstating it but not that much, lots of people are very focused on making Spinnaker more efficient it's like the urban method that Spinnaker is very expensive to operate I think we're going to need to have that yeah so that's the advantage so they have been using it at a very large scale and so with Spinnaker it's that's the reliability kind of thing it has the mileage in the field deploying different environments the cost is being addressed but you get that right I guess there are lots of different dimensions there's human cost to it but I'm really talking about computing resources if you if you it sort of depends how many places you're asking Spinnaker to deploy to if you're only asking him to deploy to let's say one AWS account and that's an AWS account that doesn't have very many auto scanning groups or security groups or images or any of the AWS resources if there aren't that many and Spinnaker doesn't have to work that hard to understand what's going on there but if there are 10,000 accounts and each of them have thousands of those AWS objects Spinnaker works very hard to keep track of them and sometimes it doesn't necessarily do it in the most efficient way and when it doesn't people can wake up at night and work so we're getting there and it is these issues are sort of in Spinnaker terms what we say cloud provider specific so the code that handles that for AWS account is different than the code that handles it for Kubernetes accounts and for all other different kinds of cloud provider so if there's a specific thing that's inefficient you can home in on that thing and fix it they're really they're checking and then they also have their vendors and say hey if this thing works then you should be trying it so you might also read the code for Kubernetes Spinnaker and the difference is they're just reading it and you understand the difference for the total version of Spinnaker and we also point this into the other but so that's all we're going to do there's nothing I can do about it but it's like to manage all the things I still need to come to this much this much can't work I'm sure we'll do that there's no way to do that Spinnaker does two things one is it keeps track of what's going to be running any time you're going to run operations you can come in and say what is the status of your deployment in this cluster currently so that's part of the way it does the periodic regression of what's currently running without your deployments stores all the memory so that's the cost we have a large number of these clusters that you're deployment to and with the different kind of resources of these clusters you grow the but he's going to talk tomorrow I'll get some of the efficiencies that you brought you know pipeline complexity issues too that we are all super super there's a pipeline complexity issue that we need to pipeline to be passed in the future I think it's very nice to hear actually I want this to be a positive comment I may not have my words organized what ends up happening I think it happened before I got there you guys probably see it in a bunch of your customers that you know the POC happens and somebody makes a pipeline or makes ten pipelines and if you're lucky things go forward people get very excited but you still got these pipelines and now somebody is going to come up with an abstraction layer to build pipelines for other teams but these original pipelines are probably still around and they probably have to keep working and sometimes you wish you could change them and pipelines get complicated even with the best intentions in the world and this can be a hard problem you know there are blue-grey deployments and you know inception of things about spinnaker upgrading spinnaker and these are all classic CD problems I feel like we're good at solving we know how to talk about like changing pipelines and people start to get hurt feelings you know and they can be hard to change sorry, do you mind using the bike? sorry the short of what I was saying is that it gets complicated because you can have pipelines they call pipelines you can accidentally input loops and through passive conduct you can just run out of memory and the failure itself will probably quickly cause what happens because the orchestrators are unaware that you've created this problem and then the next orchestrators will just try to pick up a task that is still there not realizing that it's actually it's much better so that's the catch problem that is always in the back of annoying, that's the main reason and finally what is the one track that's going on and hopefully recognize that but you can also teach and train and so we've worked with the students who have been responsible for the pipeline and they say these are the ones specifically that handle these problems talk to us kind of and there's also a reason how do you overcome those significations that you hear and how do you progress we have so many monitors in place we do a lot of monitoring of our monitors and we also as we see that we have a lot of people who have actions in your organization are you going to be the central team that provides the tools to accept the tools that you use? yeah yeah, we're going to define the path and from there we're going to try to get the engineering developers to come along with us in the CI portion of it but yeah our team is going to be kind of the core to make that happen now we're just going to chair a pickup technologist which one is good for each one but it's kind of hard because it's so many out there and you can do a PLC you know you're going to go the whole year so we kind of centralize any experience so I'm just going to do some push backs from the engineers but I'm sure you'll have to sit with them and understand what they are but one of the things that I'll speak is like do we have any training is there a place you could go or take a class? I know that our money provides training I think it does often provide training as well yeah, we provide training on need basis so you don't actually have training sessions going on but if it is in need then we pay off some input and we'll provide training we'll provide training yeah like some blog posts and stuff and videos so maybe it depends on the particular topic you're interested in but yeah, these are good resources and have you found your way to spit it first like it? yeah, I mean I'm David Byron send me a note on Slack we'll try to get you connected with the right folks and what I've seen is there's different trainings but you all would need operating spit and grit training and then your developers need to know how to use spit and grit itself and that's an important piece of that puzzle too where are you based by the way? in San Francisco oh great almost all of us Cam's in San Diego in San Diego but aren't we just technically based in San Diego or the Bay? yeah so you guys are all in the Bay? I live in the I'm in the East Bay in San Diego yeah, I'm in the just the Alps in San Francisco alright, is there any tricks? perfect, that's good yeah, I know that there's a meetup publication in the Bay area too so if you guys are ever around I don't know if there's step off specific ones there used to be pre-COVID but it's been a while I know in San Diego we are trying to revive the meetup community it's been relatively successful but meetups are kind of a cool way of getting the word out there yeah, I was just I don't know if this is going to end up being a rant or a helpful one I think in some ways when I wish there were best practices documents out there and people could say, oh this is how you write a pipeline or this is how you don't write a pipeline and we maybe don't have as many of those as we wish and I think I end up spending more of my time on the implementation side maybe making it harder to make mistakes so that you sort of can do whatever you want or at least try whatever you want and Spinnaker will tell you when you're doing something that doesn't work and feel like we may have some slightly deep waters to walk in about adjusting the way Spinnaker does authentication to make rate limiting easier and to make like quotas for the various ways that Spinnaker slices and vices this work so that, you know, most of the neighbors are not always servicing things and maybe these are all things that Spinnaker cares about and if you have smaller instances of Spinnaker or more dedicated instances of Spinnaker you don't have to care about we do wish we had better meeting and then the kind of thing and, you know, more open source dashboards dashboards actually, I'm using no progress at that so if you're not, you won't be starting from scratch but it's not as easy as I would as I wish it were yeah, I suppose it's a lot better than what you used to you'll get Spinnaker came out in 2015 as it was open there was an interval project at Netflix during that time and it was an open source leader I believe it was that same year after Google joined the project I think some of the commits the kid history goes back yeah, so it's a pretty mature product and one thing that's really interesting in this there's, you know, companies are large and they have a lot of footprint not everyone has Kubernetes only as much as maybe people might like like that idea, there's industry trends that are pointing that Kubernetes only is not necessarily the right way to go and Spinnaker's been around for four Kubernetes it works great with Kubernetes but it also works great with not Kubernetes too so if you're a large organization or maybe you acquire a company or something and you have to handle them onto your DevOps platform if they're not on Kubernetes, that's going to be a challenge and now you're trying to integrate them while moving them to Kubernetes and that's just kind of the main thing so if you're able to meet your customers developers and them when they're at you can allow them to use it right towards the job essentially I forgot how long we're supposed to go it's 10 past 12 I had till 12 I had till 12 I had till 12 11, 3 to 12 should have well I think this has been a great conversation thank you all for joining today any any closing remarks that anyone would like to make I mean it's on a schedule so I think we need to show up at least and see see if it's here I mean I think that table with some interesting Kubernetes stuff it's another place we're in we'd like to build some momentum let's see I think for as much as this might have felt like it was in the weeds that has the risk of getting even further in the weeds very, very specific details about it's been good work so now it's been offered to Kubernetes but you know at the end of it all these are the things that matter and some of it doesn't work we got to check if we want to get in there get our hands ready and fix it and see yeah yeah so join us later today at the Clownsig I guess if you're watching the recording I don't expect that video online as well thanks pretty much, catch you all later great the side of the moment I'm sure I should I don't know if you've reached the reading voices phase of one thank you thank you thank you guys for taking the session to answer my questions it was helpful are you are you I think yeah yeah we're getting to have Mike today we've got him next in so he's just we did a presentation but that would be part of the video but although we created this Ethos platform which is kind of a hybrid gigs of minotism and brazier and so on at the end we have all the standard eyes so all the teams on the low left so but it's very kind of locked down you can use it as an option it's good for the whole company So you're going to say the real or kind of you? Well, those were part of the technical oversight. So we informed about your day a little bit less than a year ago. We can find the story of Canadian technical oversight in the minds of a bit more natural. Yeah. This is the project of the report. We're going to be talking about legal and spirit review that happens in the set up of the news where we have a full request to try to decide what artists are going to be and how or what are the things that we're trying to do that are going to be done in the future. You've got to see the spend of the project. These people are so nice to the community. I think the guys... And lots of people and communities and people are worried about this kind of thing. I think, too, that people have a writing problem It is, it is a, it is a large scale. It is, it is a, it is a, it is, it is pretty cool. It is, it is, it is, it is pretty cool. I like that I saw this again. It could see you, but they're not using the mic, so I recorded again. Which is still a little bit hard to handle. I just, I saw a lot of, for now a lot of people like that. Yeah, exactly. I mean, it was more than just coding, right? And then Netflix stepped back. And Google stepped back. They were leading a lot of projects, and certainly keeping a lot of the build-up structure alive. I don't know. I'm asking this, but I'm not sure if the company would have to put some cash there to help us expand that. We're like, it's tight. Yeah, I mean, I don't know, this is part of a bigger build-up. At some point, especially if you have, I don't know what technology it uses to make the products that it sells, but if you have a bunch of drop-in golfers on your staff, it's going to be just a bunch of drop-in. So at some point, like, has it worked? You can call already, and then pay them to answer you, or call us, and then pay them to answer you, or you can look into the code. And at some point, the problem's getting hardly enough that you look at the questions. But at the beginning, it doesn't make sense to invest in that, and you wouldn't necessarily know what you're looking at. So listen to what I'm saying, it doesn't start with you looking at it, and then you look at it. Yeah, people with these troops, artifacts, five-ringes, and so on, it's for the community, people that hold them, and use them all the time. At some point, you'll start building around, and at some point, you'll start looking at it. Or make plugins. Yes, yes. 15,000 customers this week, they just got added, and you don't have time to do that. But you don't think that people just say, you know, this doesn't come off. Yes. Yeah, thanks so much. Can I have a chat?