 Greetings. Good morning. Greetings. All right. Um, I don't know today a special holiday. It is a bank holiday in the At least what UK I know. Oh, all right. Oh, I think it was been called I'm trying to figure. Oh, yeah, give me a second here. I was just trying to look. Well, speaking of holidays next Monday. Is Yeah, Labor Day. It is. I think we should. Well, my suggestion is we cancel the worker ahead of time and let people know. Yeah, it looks like it was UK. I'm not sure if it's Philippines or what, but there's a. I'm trying to figure it out. Yeah, I think that's That looks like that's Philippines as well. I don't know. Anyway, um, I didn't know that I have some colleagues out today in the UK. So, uh, I agree, probably should just cancel that one and in advance. I did intend to Last week, Victor was on vacation and I did want to set something up for us to continue on the single. Concern draft. I didn't get anything scheduled there. So I will reach out to him. Yeah, I'll reach out. I'll try to get that squared away and then I'll share that I'll do like I've said before in the past I'll just share it on the slack channel if anyone wants to join. I'm up for this afternoon. After 3pm central 4pm Eastern. Same thing. Tuesday, 3pm central. Okay. Those should give. Let's see if either of those work. Okay. Um, I think, Victor. Let's see. Do we have any, we have a PR. What is this? This is something I updated. Oh, this is, we didn't merge this. These are just small. You know what these definitions are. I'm going to. I think these could be merged. Lucina, do you mind giving these a. Plus one thumbs up. This is in the proposal document. What is this? And here's a. Had the definitions for motivation and goals. Good morning. Yeah, I'll take a look. Who else joined. Hey, Drew. Do a quick review of this. Okay. Yeah. Yeah. I'll take a look at the PR, please. Just two lines. I'd like to get this one merged. It's very simple. Yeah, I can. Get in there as well. Didn't I already. You've already done it. Yeah, I thought so. I was just, I was like, I recognize this from the draft. Document they're working on. So. Yeah. Yeah, I can take a look. I'm moving on. I don't think there's any new issues. I think right now we're. Need to close out some of the existing. Nothing new. So. You're the same primary focus just getting this document then. Yeah. Have you seen if we have any new comments since last. I have not looked. I had a little busy week last week. All right, let's see if anyone's added anything. So I think we're good on the summary. They're from. Hey. Hey. I don't know if you'll saw, but I dropped a pull request. And please review. This PR. It's very small about. The definitions on goal and motivation going into the proposal. The proposals doc. And then we can merge that pull request. We've moved on to looking at the. One process per category. Remove that later. So we're trying to move stuff over between goals and motivation to either direction. So. This is non goal. This should really be. Under the. The goal section and not the motivation. Right. Is there all the out of scope stuff. I think we have enough examples that I can delete this. I'm going to remove these. Start cleaning it up. Oh, we moved this. So we last time we took some time to move this. I'm talking about complexity of the tightly coupled CNS. Test coverage. Different teams block in each other if they're tightly coupled. So then we move the. Goal to try to solve that. And this is a motivation. So some of these are more goal versus motivation. So if we could get the motivation behind it. Oh, maybe making this a little CNS. Okay. So CNS. CNS consumers. Challenge with complex and how they come to CNS. Okay. This becomes more wordy. If you don't read the first part. If you read this, then it goes with each of these. Yeah, I was. Yeah. So turning this into a challenge. CNS challenge with complex. So. Scaling specific processes. Versus. All. Containers and processes at once. Something I'm trying to figure out how to say this, but. Do you all understand this one? If it's tightly coupled then. Then you can't do that. Yeah, you end up scale just what you want. You end up scaling it all as a monolith. Yep. That's why with the efficiency, the efficiently, you know, doing that, you know, that same type of efficiently, we're saying you end up doing everything. Yeah. Well, I think in the, it's kind of like with the, I think the, the idea behind that is like. If you're running an HTTP server, it would scale up more child HTTP process over under the parent process. Not necessarily like branching off new parent processes. We're talking about the bad stuff that happened. Okay. Sorry. So the, yeah, the challenges that people have if they don't do it properly. I had to look away for a second. I didn't know what on section. How's that for the. A problem. Not this bottom part. I'm moving that over, but this. A challenge and a problem. That we don't like. And then this is a motivation. This is the goal. So the motivation is. We have reduced efficiency of resources. And increase scaling time since you're doing everything at once rather than specific smaller components. Your response time, increased scaling time, increased scaling. I want to say and response time. So this is the problem. Our motivation on why, where you're saying this is a good idea. And that our goal in. One process. Type per container is. This. To be more efficient. Shouldn't be the goal to scale. To reduce it. The response time in this case. The response time. Yeah. Basically the problem that you're putting in the motivation is like. Basically. It's getting a monolithic. Resource. In the response time, right? Yeah. So. So shouldn't be the goal to reduce the response time. Yes, that's what this part is. Okay, I got you when loading. So scaling specific products. So here's this is the motivation. I'm going to put it like this. So it makes more sense. So the motivation is this line. This is the bad thing that we don't like. So we're calling out the problems and pain points challenges. That people are facing. And then what are we trying to do to solve it? So here's the goal. So. The problem is a monolith is going to have. An efficient use of resources and they're going to have a. Slow response time. And our goal is to more efficiently use resources. When, when you need to scale. And we want to decrease the response time. Okay. Does this look right. Go ahead. No, no, no. The only thing that I'm now confused is between. And solution. So probably it's pretty much the same. Well, the solution that we're giving is all the same. It's. One process type for container. And the goal for that solution. Is what we're writing here. So right now all we have is one solution. That's what the best, the best practice is the solution. And we're matching that best practice up. We're matching that solution. To its challenges and what we expect it to provide. I think part of our problem is all of this is separated. And it might actually be. It may. You know, help. Okay. So it's possible that just because we dig into it, it becomes more confusing, but we could think about the whole format. If it would be better to have like motivation and goal. And every time you put something, you would want to have, have it all together. So your challenge and then what you're planning on to do. But right now what we've been doing is we would take that goal. And we're going to add it. Yeah. Because not all the motivation has goals, right? Yeah. And that's why we were not putting it all together. We were trying to separate them out. How's that. For the goal. Yeah, that's fine. Yeah, it works for me. So some of these look like they could be common goals. Increase the uses of Kubernetes orchestration capabilities with CNS. From a CNF consumer operator standpoint, that means that the operations team is going to be managing more things in the same way across CNS. So if they're all doing leveraging the Kubernetes orchestration capabilities, then they're going to know how to manage them. They're not going to have any issues, whatever. But that's also true for the CNF developers. If they're leveraging the capabilities of the, of the platform instead of trying to run a monolith and do their own orchestration within a large complex CNF. Then you can focus on other things. You leverage the platform's capability and then you focus your development elsewhere. So this in my mind is a common goal. We could have something like this shared goals. Same thing for if you're aligning with microservice architectural practices. If all your CNFs are following my service practices, then as a consumer, the operator, your whole team can be trained to manage the, the CNFs using those patterns and practices rather than having lots of different ones. And then similarly, the CNF developer would be better utilizing those practices when they're doing development and whatever else. So I'm going to speak up if y'all object or. Well, I was thinking to include the role in the phrase, like for example, that one reduce the attack surface and just associated with the security engineers. So, for example, saying like security engineers can reduce the attack surface area, multiple process, something like that. Like I just, just including the role and that way we can remove all the roles from, from all the sections. Is that all under one big goal? Yes. That's another option. Let's see how that looks. What do we call this? I don't know if it's a cycle management team or something. Yeah. Maybe we don't have to call it security team for that one. It's just generic. Whoever wants to do it. Okay. And I don't know that we need to say the operations team. That's just, who is it? You just get support scaling individual processes. Well, in that one, I'm just wondering about development cycle cycles. Yeah. Yeah. It's obvious that we are assuming that the development team. How about that? I'm not sure. Yeah, it's redundant to, to put multiple times development. How about that? I think this is good for everybody. I'm going to move the scaling. I don't know if we need to do it like that or not, but actually I'll just put it after. So one is just figuring out how do you want to scale it? Well, I was thinking two things. It's actually scaling it. So, so in general, I guess that we, what we're doing is simply for simplify the. Some things. Yeah. Well, one of the things is like, I don't know if we need to do it like that or not, but actually I'll just put it after. Yeah. Well, one of the things is like the troubleshooting and the ability of the lock output. And the other thing is simplifying the scaling process. I don't know how I should, we should put it all together like that. So all of them are just simplifying like. Yeah, I'll just put it here today. Consider combining. And simplifying. I'm going to move back up so we don't get stuck. Wow, we got so much. Okay, motivation. Large surface area for security attacks. Okay. This is a challenge. Difficult to identify where the bugs are opaque view of the communication upgrade causing interruption with many services. So these are all challenges. All right. I'm going to leave these separate. I think it looks good up here in the motivation. And then the goals. To me looks okay. All merged together. Now these are like benefits. So it should probably be goals. The best practice benefits. Seen if consumers and following ways. Okay. Improved security through isolation. I don't know. How this is relevant. Use the container name system. So I guess they are. If you split the processes. And you don't have the same process type in the same container. So now you've split them into different pods. And their own container. Well, then they are somewhat isolated. Of course you could have problems within the entire scene of namespace. So I guess they are if you split the processes. And you don't have the same process type in the same container. And you don't have the same process type in the same container. So you can split them into different containers. But I don't see enough namespace if it has multiple pods or whatever. But just separating them into different containers. Is going to give some isolation. Then separate them into different pods more. But this would be a. A goal. I'm going to move that down to goals. Not. Unless we're going to say something like. In. Let's see. Security. Bugs in one. Process type. Effect. All of their. Processes. In the same container. Does that make sense as a problem when you have a bunch of different processes. Together. And if, if one of those, let's just call it different applications to make it simpler. If you have a bunch of different applications in one container. And then you have a security bug in one of those applications. Well, then it's now. Caused a security problem for all the other applications. Is that clear? Let's do the. Like Apache and. Postgres. Postgres SQL. If you have those together and there's an Apache security bug, then all of a sudden your database is. At risk, but if you have your Postgres database in a separate. Container. Then it's. At less of a risk. That's what this is trying to say. So, so what about just instead of books? Vulnerability. In general. Security issues. Or when I read, when I read this. My eyes are messing up. Vulnerability is in one process. And then we have. The goals. Let's see reducing attack. Okay. Then we do that. Now we have this. So this ties in with that last one. That's a goal. All right. Okay. Now we've. Done that. Oops. Let's take it away. I think this goal is already there. I'm going to just put it here. And we know that we need to merge all of this. There we go. We can merge all of that scaling. Okay. This one is about. A goal. Reasing about. The log messages and all of that stuff. And then we have. That and. One of these. Difficulty to identify components. A peak view of companion communication. So we don't have anything specific about logs. Oh, I guess we could say component. Log messages. Are more complex. Because. They. Are from many. Different sources. Source. Yeah, I was going to say process type sources. Instead of a single. Process type. All right. And really, I guess I'm saying log messages. From a container. So we're not saying the whole CNF because the whole CNF, you know, maybe there's something. Or the product or whatever. I'm going to say that. So you have a product that maybe have multiple components and everything, but this, the product is broken up into micro services. In each of those micro services, you know, as a single process type in a container. Well, now the product may have like a. A label and all of it ends up coming together. And you want to make it where you're saying, oh, this all from that one product. But then when you're debugging, you know, oh, each container only has one process type. Versus if it has lots of different processes and you don't know what, what was it? So that one would be. This. This goal. I'm going to put that there for now. Oh, this one. I see something there. That's a very simple version of this. That we didn't finish. So, so regarding locks. There are a lot of local aggregator tools. You know, you don't need to rely on those tools, right? Like, I mean. You have like a. If you want to troubleshoot something. So, so you can go directly to the container and you will see only one unique. Log instead of like. Having to correlate all this. Yeah. Instead of writing or like, you know, things in many, many places or, or, or like, I mean, mostly like this implication, right? Like. The fact that you have multiple locks. It's not affecting the. The, the operation of the container science. Yeah. The CNF, right? So. What I mean is it's. You can, you can get the similar. Benefit, but you have to rely on. Additional tool, like a log aggregator to. To collect all the logs and. Put in a single centralized place and. You know what I mean. Yeah, you may. Decide the use. There are a lot of. I mean. Yeah, you may. Decide the use. There may be some type of tool that helps you to search through stuff and. Do some type of analysis or whatever. Yes. Rather than depending on that tool being. Really. Intelligent in its capability to. Get together and label everything from. Oh, this is a multiprocess. Oh, these are different. Apache. This is Postgres, even though it's in the same container. All of it becomes simpler. If you. Follow the same standard. So you. Some of those tools that you're talking about. Have. Become very advanced in their capabilities because. Everybody was following standards. So they just had to add more and more. Ability to. Identify things. Right. But if you are following like just the very first. Standard of, you know, this is related to 12 factor apps, but. Everybody print their logs to standard out, start there. You know, then immediately. You're going to have. You're going to have the. Container. Plus a log. And then if you have. One process per container, then we immediately know that. Okay, that's the whatever process. That's the Apache container. So all logs from that are the Apache logs. That's the Apache. Now, whether you end up using something external. And you're pulling it in and have like some. External monitoring and observability system that may like give alerts and page people and everything else. That's something else, but you could. When you do that alert, you could point everybody right to this container. This container gave the alert. And then you know, there's no other processes, no database, there's nothing else. So the whole thing becomes easier. So now someone like the admin or someone jumps in and goes, oh, yep, we need to restart those. They're having problems or they go, oh, let's escalate that to the vendor. And it points it right down and, you know, they have VPN. They can go in and connect and work on it, whatever it is. And that's where the procedure becomes nicer because they follow the standard. Yep. Boom. All right. So this is another one that we need to, you know, update and simplify. We had a very simple version of this, but probably need to take in some of these other things. Let's see. So we only have a couple more here. Unprecedented. So this is a review of improvement. Life. So management deploy ability when a process binary or poison container. Of course, grain relative to the rate of change of the binary container. But fine grained relative to the rate of change of other processes and their dependencies. Yeah. So then redeployment can be happened based on. And like a container level versus are all the processes and libraries within the container going to work out, are we going to have problems? But again, this would be a goal. I think I could just move this down and we can look at updating or not. I feel like this is going to be related to like upgrading. This is related. So we talk about the orchestration part and increase the use. So there's something here with the orchestration. So maybe those need to be merged. This is becomes, I don't know if this is an operator or not. Now we're down to something. I'm going to just put it in here and then we can move it. Having multiple processes requires orchestration. All right, I feel like that one needs to be, we need something there. Okay, this is a goal. And it's somewhat related to this loosely coupled components. And then we already have a testing. Test coverage, but we have it down here. I'm just going to merge it in. We can simplify later, but I think we, this is, I think I've moved all the goals down. So unless we think of more motivations like this one that I just added, then we may be good with the motivations and then we clean up the goals section. And we should be good. Then we can move on to the proposal, which we thought would be more straightforward. Based on all of this, what are we proposing? We have a summary of what we're proposing. And then we actually put it all the way forward. So under motivation, we could do the same thing that we did in goals and just merge it all and not have different sections. And we can see that someone that's a operator or someone that's developer may want to go to the different sections, but we may be breaking it down more than needed. Maybe they just look and find the bullet points that matter to them. We don't worry about making sections. Yeah, I'd be okay with that. All right. You can go accept those Oliver or Victor either one. I suggest that it suggested. And then I think we're done with motivation. All right. Okay, so we got to simplify goals. But we feel good about what they are, we just need to combine them. And if you're thinking more non goals out of scope things we can add them, then add the proposal. We've already said the workload context is all pod types. We should add some user stories here or add them directly in line or add them as new user stories that are shared for anyone to use. This is where we say it's okay to have multiple processes. And then we have some references we could add more. If there's any alternative options. I guess it's going to be first version is in September. All right. Well, we're close to the top hour let's stop here and if anyone's up for a working session this week. Oliver is going to post a message into the Slack channel. And then we can get together. I think we can have a PR. For next Monday's review. And we'll take it around next Monday is holiday. Oh, that's right. Yeah, we're going to we're planning to take off anyone object to that. Let's hear plus one or on cancel. Object. You agree, Victor. Yes. All right. Who else do we have on the phone. Drew. I won't be on next Monday. All right. All right, we'll cancel them. I don't give us more time, but I'd like to get the PR and we can tag as many people as possible and get that one. The PR reviewed for this. Best practice and move on to the next. Let's take something from nephew. Something on configuration management or the get ops practices. Some of the stuff that's built into it deployment. Simplifying orchestration. It's really replacing like. Mano and. A lot of the other stuff out there that's more complex. With a simpler way of doing things. Yeah, updated modern fitting 12 factor app and everything else you kind of think so. You're involved let's think about that see if we can find something. I guess that the, the major. Differentiary in the fear nephew is. The scale like everything is sinking in a scale like. To deploy thousands of CNF. So that's probably the major difference. Anyway, I guess, regarding to your open PR. Putting notes on. Motivation and. What else like a gold. I think that you can merge it. So. At least you, you have. The current PR we can merge. Yeah. Or do you need more. Approval. No, that's good. I'm going to merge it. I will try to fix all the LinkedIn. Or my issues later. All right, thank you. We'll get one started for this. This week, I'd like to get the. One process type practice. If I'd like to put, put a PR in for that. Yeah. I will send a message over and you guys let me know. I'll mainly direct it to Victor and to you Taylor and see if we can't set up, you know, at least the session today, if not today tomorrow, you know. I'll fire off a suggestion. And if that works for you guys, we'll meet. Otherwise we'll just move it. So we figure something out and then I'll share. I'll share in case there's anyone else who wants to join. I'll share that on the slide channel. Sounds great. Yeah. Thank you. Yeah, no problem. All right. Thanks everyone. Bye bye. Bye.