From 4bdd828493b890367168aa8be62c9500344b1c99 Mon Sep 17 00:00:00 2001 From: Drinkwater Date: Thu, 22 Aug 2024 21:36:01 -0400 Subject: [PATCH] Update 017-2024-08-20.md added transcript Chore: add recording link Signed-off-by: Drinkwater --- sig-clients/meetings/017-2024-08-20.md | 112 ++++++++++++++++++++++++- 1 file changed, 111 insertions(+), 1 deletion(-) diff --git a/sig-clients/meetings/017-2024-08-20.md b/sig-clients/meetings/017-2024-08-20.md index 3c446abb..32a70d51 100644 --- a/sig-clients/meetings/017-2024-08-20.md +++ b/sig-clients/meetings/017-2024-08-20.md @@ -188,4 +188,114 @@ - **Dominikus Brian:** - Test different GPU and storage options and provide feedback. -## Transcript +# **Transcript** + +Tyler Wright: All right, welcome everybody to Sid clients by monthly meeting. During this special interest group for clients the group here gets together to discuss anything that's going on with any clients being built on Obviously one of the first clients that was built on top of Akash was Cloud most which is now again update on Console being that again. It's the leading deploy tool for how people Deploy to + +Tyler Wright: I know there's some folks that usually join us that can't make it here that talk about terraform and also the gold client libraries. I know there's going to be a fair amount of work that goes on there as we move into the cosmos SDK 47 upgrade. So again when that test net starts some of the themes behind terraform providers as well as the go client libraries, we'll have again some additional work cut out for them in terms of making updates to those clients ahead of the network upgrade. So again, that work already been mapped out and discussed. I'm sure we'll have Some further discussion between meetings if not the next meeting. + +Tyler Wright: As a last couple of months, we've changed a Cadence where this used to be a a monthly meeting just because again, there's a number of other places including the steering committee the product and Engineering roadmap where you can get updates from the Consulting all the work that they've been doing on the client side. Again, there's a number of other teams that are building various clients on top of Akash or having the past getting them to join this meeting sometimes difficult because again, they have other responsibilities so we look for updates from anybody that's building any clients on top of Akash to again, feel free to chime in to the Sig clients Discord channel to give updates. If you're cannot make these meetings or if you have ideas for things that you want to add to console or things that you want to add to the user experience and maybe a form of another client again, always + +Tyler Wright: Looking for folks that are building the cost to join these conversations get involved talk about problems that they're facing Etc. + +Tyler Wright: you can also track many of the past meetings in the community repo. There's a sick clients folder where you can track again some of the stuff going on. I know there's been a big plan that was discussed at the last meeting about everything happening with console 2.0. So again, we'll just get update on some of the work on console and then if anybody has any questions or any agenda items that they want to talk about feel free to drop those in the chat and then we can talk about those here today. + +Tyler Wright: But I do want to hand it over to the console team again. We take a part of this agenda to get updates on all the work that the console team is doing across a number of different initiatives. We've talked previously about some of the work related to the indexer during this again analytics by monthly meeting. So we don't have to necessary talk about that here, but I want to hand it over to their Consulting. Maybe Max Anil to talk a little bit about some of the advancements over the last couple of months and look ahead of maybe what's to come? + +Maxime Beauchamp: Sure Anil you want to go? I can go. + +Anil Murty: Sure, or you can go first. + +Maxime Beauchamp: But yeah, I will wait in the past month and there's been a lot of effort and progress towards enabling managed I will on console. so users will soon be able to make complete abstraction the Wallets on chain. So everything will be so a lot smoother for users. So that is almost. The MVP is done, and we're almost there to release it on production. + +Maxime Beauchamp: We've been working in hard as well on fixing bugs that customers are facing right now. There's an more usage more bugs that we find and we work hard on fixing these and just improving in the overall experience for her console users. + + +### 00:05:00 + +Anil Murty: So I know I can jump in a little bit. But yeah as Max said the big Focus for the core team has been to implement the first phase of managed wallets Plus Credit Card base payments in console. The very first phase of that is basically implementing what's called a trial wallet. And so the idea here, is that a user that comes into console? Should not have to have a wallet. She not even have to log in and can just try console without having to literally do any sort of sign up or login. and then once they try it and maybe do one or two deployments then they can choose to obtain to either a custodial wallet solution where we manage your wallet for them and they play using credit cards or a self-cously you want it where the experiences exactly the way it is right now in console. + +Anil Murty: So the very first phase of that is the trial wallet portion and that portion is currently in beta test within the ocl codeine with an overclock lab score team. So we're going to spend a few days later testing it outside of the engineering team within the broader or Cloud lab scheme here. And I'm hopeful that in the next few days, we'll be able to open up the beta for insiders as well as for. Potentially a upcoming hackathon that's going to be happening in India. So if you don't find any major issues with the trial experience this week, we're hoping we can actually use this for the hackathon. What's happening in India? So that's the big update. Other than that, if you look at the project board that Tyler shared some of the other big updates you'll see that Max talked about + +Anil Murty: which are in the bucket of enhancements as well as bug fixes. + +Anil Murty: some of the things we have done is we have carved out all the shared packages into a separate section inside the code base. And the reason we have done this is so that we can build things more in parallel across the team particularly as the pretor team is starting to work on the fronting new contained user interface for provider console, which will be basically a portion of console that is going to sit inside consort network, but it's going to be basically a new front-end for spinning up providers. So all that work can happen in parallel. Now that the shared packages are shared out and we can reuse components across the deployment console as well as the provider console. So that's issue number 230 on the list and then one of the other things that was added, I think since the last time we met here was persist in storage of being tracked now in the stats app as well as in So when you went to Stats or to console providers + +Anil Murty: stories that you'd see there was only ephemeral storage and in the last few months, you've added a bunch of providers with a lot of persistent storage and it was important for us to start tracking all that as well. So even though it's not broken up into its own thing into its own donut chart the storage overall storage now reflects both FML storage as well as persistent storage. So if you notice a little spike in the stories, that's what that is. and then the third significant thing worth calling out is we have also updated What's called the exponential back of algorithm for a new providers that are coming on to the network. So what this does is basically + +Anil Murty: improves the user experience of how fast a provider shows up on the network while it being set up and also if for some reason a provider failed to check before the exponential back of algorithm would back off to a day or more and that would sometimes cause the provider to show up offline even though it's technically online. And so that was a pretty big fix that Maxi made and that's one of the board right there. There's other things that are happening on the cd side. I don't see the human HQ team on here, but there's a bunch of progress that's happening on implementing. Basically what I call a versatile style experience within console and push and the human HQ team and as we sync up pretty regularly on those between these meetings and the pretty close to having a solution that'll work really beautifully within console where the user will be able to come in connect to a GitHub or big Market or lab repository be able to pull in their + +Anil Murty: application code and then have the backend basically the software build it for them deployed onto console and just give them a main point so that way they don't have to authorize their application. They don't have to write an sdl none of that. And then also basically Trigger for being able to redeploy the application every time there's a new pull request into the Repository. So that exciting stuff is coming as well. And in the context of that they're also going to be looking at some testing out some logging improvements of sending console logs to a third party logging service and then cleaning up the logs and thinking in the back into console. So I'm hopeful that those changes will get more into console before the next time we come here if not much sooner than that. So I think the crcd work combined with the trial wallets and the custodial payment solution is going to be a huge driver of option on Akash in the coming months. + + +### 00:10:00 + +Tyler Wright: You Max Ben. Thank you Anil for those updates. Does anyone have any questions on all the work that the client team has been doing in terms of console and again bringing members of the community to also help? + +Tyler Wright: Cool as Anil mentioned as well there with all the work that jigar and Duval have been doing on bringing the Prater app into console, but we'll talk a little bit about that a little more detail there in the state providers call happening tomorrow. But again just another shout out to all the work that they're doing to unify the console experience. Go ahead dummy. + +Dominikus Brian: Hi everyone. Thanks for the update guys. It's very great. I'm in this console is becoming more and more. I mean, I have just tried several of one concern with this plane Linux with SSH so I tried it several other days, I mean Using trying to SSS there. I love this, generate key function. that's a perfect thing and then everything is just there. You can download it. Then I can just copy the public key and stuff. That's as easy as it gets. I mean, you don't even need to generate it yourself and that's nice. I love it but there is some issue with choosing this providers. So most of the times depending on the configuration I asked for the building is not always available. + +Dominikus Brian: And then I need to sort of manually downgrade instead of getting an options to if say I'm not sure which one is not affordable. Maybe I'm asking too much of memory if it's asking too much of something. so the GPU is already pretty General I'm asking only for NVIDIA and then I'm not sure whether it's the storage too big and other things. I need to just keep trying several times. I think this is a things that can be improve. So either doesn't show a bidding for that. I will be shown some beading for lower, + +Dominikus Brian: Great for instance. I asked for 300. There is an options like 300 is not available. How about 200? And in this bidding is also there you don't need to show everything but I mean the closest one to that will be a nice thing because sometimes it's a bit irritating to close all the deployments bidding. And then I go again trying to with lower quality one. I mean that's a thing. I mean that maybe can help accelerate where people feel it's like failing other it's just a little bit of requirements. So yeah, that's one thing and another thing with this connections when connecting to that sometimes it's not stable. I'm not sure why + + +### 00:15:00 + +Anil Murty: the nice words on the SSH option or the VM option basically, but yeah, we have heard very similar things from a lot of other users in customers as well for the longest time people would struggle with setting up their SSH from their own terminal and giving them this option and then also having the references deal in Akash is going a long way these days in terms of getting those customers onboarded. So we've heard very similar things where people are This is very much like the experience where we get to see in front if not, ec2 itself. So that's great. I want to address the two things that you mentioned or the I think the first thing you mentioned was something around. Not being able to get bids and not being able to sometimes adjust your parameters. So I think + +Anil Murty: there's a couple of things there and I want to make sure I capture your feedback. So one is in two places in the rent GPU page as well as in the page. What we were doing was we were defaulting the ukt value to 1000 and that filters off some of the providers that are more expensive some of the G2 providers are more expensive and so as we speak Max B's actually going to make this change. In fact here request open right now for updating all the awesome Akash SDS as well. And that will basically up this value from 1000 to 10,000. So it is an exit and so you should be getting pretty much every provider covered with that. So no provider should be getting filtered out. So that's one the second is I know that in the + +Anil Murty: as well as in the rim GPU, I'm just gonna check this again as I do this, but in both of these options, we actually don't show the noise here. That's just a design time choice that we made. we have to go with some more thing. We wanted to keep it as easy as possible with the assumption that these users are relatively less technical compared to the other users or something like that. So we hit the raw. I see from the ram GPU page as well as from the VM template, but if you think that that is something that we should consider adding that is something we can certainly look adding into that as well. It's something that we have another templates. So that's number two. There was another question that you had that I think I've kind of lost track of but just want to pause their see if you have any feedback. + +Dominikus Brian: Yeah, that's great. I mean I tried that video akt it doesn't help at the time. Maybe the other is busy. But I haven't tried much with other options with the friend GPA GPU directly without the get used to the STL. But yeah, I'll check and try the other ones and see whether it can give me more options. I'll write it back when I've tried more of them. + +Anil Murty: Yeah, that'd be great Both of them have all the options for adjusting everything. The only thing they don't have is the actual rise deal. so you should be able to change everything. It's like the GPU weapon will storage all of that but + +Dominikus Brian: Yeah, that's going to be very nice functions. Yes. + +Dominikus Brian: Sounds great. + +Tyler Wright: Perfect. Anything else go ahead. + +Dominikus Brian: And maybe it can go up together with the docker eyes. I mean you guys mentioned that in the future soon enough, we'll have this, Auto docorization from GitHub and everything. so there is the same functions that makes just describe what actually helps, giving users assessments of What's the requirement for this Dockers as the storage wise and things like that? So it will be nice if it's combined together. I guess that's all from me. Thank you. + + +### 00:20:00 + +Tyler Wright: Academic go ahead and deal. + +Anil Murty: Yeah, so on that point and I had forgotten what the third question was. Thanks for reminding Max B. So on the topic of how do we increase the likelihoods of bits being received? I actually created an issue to basically collect all the possible scenarios where that problem can happen. So if you have any other Feedback or other situations where you think this could happen or other ways of handling it feel free to comment on that issue as well. But I try to capture all the possible scenarios in + +Tyler Wright: Queen anything else on console and he wants anything anybody wants to discuss. Pardon me? + +Tyler Wright: I know it's been much discussion about awesome Akash as well. So again, please feel free to look at awesome occasion of the templates from Austin makash automatically go into console as well. So again, feel free to contribute to the awesome Akash repo if you see anything that you think might be great to dakarize and again put an awesome Akash for others in the Akash ecosystem again, feel free to contribute there as well. + +Tyler Wright: Are there any other discussion points that anyone wants to talk about related to clients on Akash? + +Tyler Wright: make sure that the notes are made available. Again. This is a monthly meeting much appreciate the updates from Max B from Anil the commentary from Domi all that is very thankful for all the work that the client team has been doing on the Shout out to whom who was also been supporting with some efforts on the CIA cicd side after getting some money from the community pool to support console. They're also been some contributors that have been getting involved in tackling some small issues related to console. So again, as more of those smaller issues get started or are made available, we'll have the core engineering team continue to work on a number of the big efforts continue to + +Tyler Wright: time to connect with the community talk about updates talk about all the work that they're doing get more people involved and again, there will be continued to be good first issues. One thing that we always ask if anybody wants to get involved in contributing to console. One of the first things to do is again, Look in the console repo and download our get console locally on your machine. That's a great first touch point to look at the again documentation and be able to run console locally. And then again, once that's done feel free to reach out to myself and Sid clients Channel or look in the issues inside the console repo and see if there's any good first issues for Community contribution. Go ahead and kneel. + +Tyler Wright: Sorry, excellent. Thank you very much Anil for sharing that and putting that together. Great again, if there's nothing else then we can certainly wrap up if there's any other notes or updates from any other teams. I'll make sure that those get added inside the stick client Discord channel in between meetings. Again, if anybody has any questions or conversation or any ideas, feel free to continue to drop those inside of the Sig clients Channel again, if it's specific to the console, there's an ecosystem console Channel as well where you can talk about any active issues again, we talk about some of that stuff during the six support. But again, there's all the issues that you can kind of track on the console side that's been dropped in chat here and yeah, I'm sure there will be a lot more to go ahead dummy. + + +### 00:25:00 + +Anil Murty: and so it would fit with AC clients even though You may be crunching data from the network and stuff like that to figure out which is a provider. + +Dominikus Brian: I Thank you for it. But yeah, because I think that's a great thing when you can have so in your say consoles a kind of help. Button, and then this agent basically, assists You actually need to do this or do that or things like that? I mean suggest you a little bit on requirements and stuff. But you need to train a lot of things maybe in the future if I have something on that are updated here. Thank you. Okay, that's it I guess. + +Tyler Wright: Mason excellent again, I'll make the notes in the recording available on very shortly again. Shout out to Gorge for all of the support with some of the notes again much appreciated the update from Max B from Anil some of the questions and comments from Domi great conversation. We'll talk a little bit more around the provider side of console during the sick providers a discussion where jigar and Devol will talk a little bit more about all the work that they've been doing relates to council 2.0. So look out for that sick providers meeting happening tomorrow. + +Tyler Wright: But yeah again much appreciated everyone's time today will reconvene in two months. But in the meantime, we have Sid clients, we have the product and Engineering roadmap that gets updated fairly frequently and look out for a number of updates in various channels on social media related to all the work that again the core team is doing as well as human HQ and many other teams that are working on console and other clients but much appreciate everyone again. Hope everyone has a great rest of the day. Thank you very much. + +Jigar Patel: Thanks. + +Maxime Beauchamp: Thanks, everyone. + +Anil Murty: Thanks, everyone. + + +### Meeting ended after 00:42:00 👋