 11 o'clock on the nose and fortunately my zoom just crashed so I rejoined and apparently I have not dropped off of the session yet so I apologize for that hopefully it realizes I'm gone here shortly or if someone can kick out the old copy of me that would be great but of course today we'll be doing the functional group update for the build team and myself at Marm will be presenting hopefully here soon I can't seem to take over unfortunately but what we can do is just get perhaps started a little bit here so as far as agenda goes what we want to cover is highlights so far in Q4 right so what we've achieved and what we hope folks are aware of and excited about across the company also some of the current challenges facing the team and how we're dealing with them and overcoming them and then since we're kind of getting close to the end of 2017 here we want to talk a little about some of the big carry goals or big ideas and focus areas that we have for 2018 that we want to try and spend our time on and deliver customers and then we'll wrap up and take some Q&A so I unfortunately don't know how to get rid of this but the link to the presentation is in and I can't claim hosts all right so the link to the presentation is in fact in the calendar rotation and so you can follow along there and again apologize for this sort of screen sharing snap do and thank you yo for linking it in the chat Martin you want to take it away here as far as the first set of our highlights great thanks Josh I think we need an admin here to kick the old you out but I don't see an admin at the moment so we'll have to do with what we can oh I see you're out so I guess you can give it a go again perfect all right let me go ahead sorry take it away yes great first of all I wanted to introduce Richard clan as a new team member he joined in October but we didn't have a functional group update in October you might have met him during the summit and Richard has already proven that he was he's a great hire he hit the crowd running and contributed a lot so a very warm welcome to Richard as for the highlights for the other highlights in Q4 today we are releasing the two so happy release day everyone we with 10 the two we are finally making EGA generally available and I have to say that it's already running on gitlab.com for over a month and knock on wood we had no problems with it yet a couple of things I wanted to note here is that Ian bound did a great job building all of this you can check the documentation that is linked there and if you take a look at the docs I have to say some of you might ask where is Josh I'm going to mute you yeah thank you I can't mute while sharing myself so okay well that's a zoom I can't mute you so you'll have to be quiet I yeah look I'll need myself here yep cool if if you take a look at the docs as I was saying you might wonder well where is the simplification well the simplification is there if you start going through the docs you'll see that you can now set up a cluster of seven machines that will make your database highly available but as always in this company we are iterating the main point the main goal for us was to have a full set of working documentation when making this generally available and our next steps are tied with the second item in the highlights here and that is getting the configuration even simpler than that if you take a look at the docs you'll notice that there is a lot of repetition there are a lot of things that you need to copy paste and basically a lot of those things are not user configuration or rather not configuration that user needs to think about it just needs to happen right because we are shaping the package that way with the roles change that is scheduled for 10.3 with PGAHA this document is going to become much much much shorter we already have this for a Redis AHA and for GL and it has proven to be much simpler and hopefully with the feedback we receive from our customers we will make this even even simpler to set up one other thing that has been proven to be very problematic for us was deprecation so whenever we introduce a configuration change or we need to remove something from the package we have a very large tail behind us which means customers and users do not upgrade that fast as we are shipping which means that six months after we do our release we get feedback sometimes that hey something is wrong or why did this break and we have no idea anymore or rather it takes a long time to figure it out we did some structure changes inside of the package and now we are able to present the precaration warnings and we already managed to catch some interesting feedback from from our customers and this has proven to be or will prove to be a very useful thing in future for us I think we can move to the challenges Joshua take it away thanks Martin so yeah so thanks for that and super excited about all those features deprecation warning as you mentioned is already proven to be helpful I think it raised some potential you know upcoming deprecation and the get host service for example that were sort of going missed with sort of the current method of having it to warn you but warn you kind of throughout the reconfigure process and now with the summer edition at the end that really kind of helps to clarify and show what they need to be aware of here as far as you know features that are you know we'll be going away at some point in the future here so that's awesome and again they can get I've easier to install is super exciting and other things that are exciting is that you know opportunity is everywhere for us get lab is you know getting increasing mindshare in the marketplace the forest away report was awesome and so we're just finding more and more folks who are wanting to partner with us and want to have a get lab and you know involved in in some portion of their efforts in their own marketplaces and so what we're seeing is that definitely a lot of inbound requests Elevron's been very busy as far as distribution channels and places to host get lab for example and you know the challenge we have a course is that we have to be pretty selective in what we spend our time on again I realize all get lab is like this but in particular you know we're having to test to kind of hold off on some of these other opportunities for a couple reasons one is you know just capacity and the great news is here that we're working hard to expand the team we have a new hire shooting in January who we are really excited about and bring some some great expertise but also you know currently in Q4 here a lot of our team resources are being directed towards a single letter right so over half the team is in fact focused on delivering one thing and that thing is in fact I'm sure everyone's aware of it the cloud needed get web effort right so what this is is aiming to embrace more cloud the practices so instead of having a single container or snow which has every single part of get lab on it for a matter most to work for us to rails to engine next to instead offering a suite of containers which contain just a single portion of get lab right so we have a container for example get a lead a container for unicorn and rails and this is a number of benefits for improved scaling because you can individually scale each of these components and resilience so if one has problems it isn't sort of taking affecting a larger set of functionality and then it can be more easily and quickly replaced with sort of the new version or or a sort of restarted pod if you will so that's kind of a major effort and a good chunk of the effort here and Marble going into more detail shortly on sort of where the effort is and what we found so far in this push the other thing we're doing is we're trying to reduce the reliance on shared storage so right now you know to get web.com and really for any customer wanting to run HA they have to have a shared storage service something like an NFS and care and feeding of that you know impose a certain maintenance overhead and a point of failure that we'd like to get rid of and so we want to try and reduce the reliance on that have fewer pets nearly NFS and sort of improved uptime and improved also just maintenance and installation as well so a lot of great benefits here and again a wanting that cloud native kind of sort of guiding star get lab and where we're going as a company and we're aiming to have an early beta here by end of year and you know it's proving to be quite a challenge as we kind of document some of what we're finding now but again still striving for that and we'll of course continue to focus on this for you know good one time after here as well to make sure it's all production ready and goes GA of course here as soon as possible with that pass it back off to Martin here talk a little bit more about the kind of engineering effort and some of the challenges and surprises that we found in thanks Josh so yes as Josh mentioned half of our team or more than half of our team is actually dedicated to working on the on the charts or cloud native get lab but I also have to say that it's not only us it's a very big engineering efforts here and it became even bigger when we tied the helm charts to get low.com GCP migration effort which means our goal of migrating to GCP now became even wider but more exciting actually where we are going to deliver the charts that our customers are going to use but we are going to live test them at scale when we move to or during our migration to GCP. One of the things I also wanted to mention is while this is an engineering wide effort it is also now being proved that this is more than only engineering it's proving to be a company wide effort we now have all of the product managers also involved from the top all the way from SITSE to everyone involved like we have a very very big involvement from everyone to get everything that is needed for us to deliver or move a GitLab to the next stage. Some of the challenges that we found which are reducing our velocity basically are I link there probably not really descriptive but the components that's definitely surprised surprised us was a GitLab shell it is just popping up everywhere it's turned out to be one component where every piece of GitLab has something to do with and entangling those dependencies there will take more effort from various engineering sites not only production or build team. We also caught some components mid-change for example Gitaly was supposed to be one of the easier things for us to ship but with their move with the Gitaly steam move to have part of their code base in written in go in their Gitaly server and part of it still being in GitLab Rails means that Gitaly is now also popping up in a couple of places they're doing an amazing job but this is definitely causing some was causing some delays and will raise some of the emergencies when we come to them and we also have a couple of unknowns for example we have no idea as of now how are we going to run database in Kubernetes I don't think many people actually do know that yet but we are working on finding this out we luckily have consultants and for now database we are considering as that which means for our GCP migration effort database will be outside of the whole helm chart delivery and as Joshua mentioned storage is one of the other things that we want to get away from I'm not going to go forward with explaining what the GCP migration effort is that is going to be presented by most likely by production team so I'm going to move this over to Joshua again for our goals as a built in for 2018 thanks Martin so again you know kind of that's gonna take us through the end of the year here as kind of again a lot of our focus and a lot of our efforts are going into that to make sure we support the migration as well as try to have a beta out that customers can try as well so that's where we'd be spending a lot of our time here as a team for 2017 or what's remaining of it which brings us into 2018 right and kind of you know where are we thinking I want to spend our time what are the areas we want to improve and the new teachers we're going to try and deliver to our customers and this is just a selection of some of the kind of most important or perhaps the most impactful here number one of course being the cloud need of GitLab offering right and taking that from the early beta state that we're hoping for and delivering it and making it really GA right and having that be a great experience for our customers and really first everything that we can having also production use the same tool set and so you know our customers can kind of run a .com like installation of GitLab just using Helm and Kubernetes along with perhaps you know like a database and things like that as well so that's super exciting a lot of engineering work going into that and really impactful I think a lot of folks as well next up though beyond that is kind of another goal which is we want to make sure that GitLab is as easy as possible to install and deploy and then also for ongoing maintenance right and we're tacking this a couple areas we saw this earlier with our deliverables here in 2017 where we're doing some work to ease the configuration of GitLab HA where we're making it more apparent of what's deprecated right so we want to try and make sure that we continue to make sure that GitLab is you know even easier to use even easier to install even easier to maintain and just continue to drive down our operating costs as well as making sure customers can try it as quickly as possible a little effort as possible so they can start to see value in there for the concepts more quickly so to that to that end we're kind of focusing on a couple things here as far as big goals we have a number of other smaller improvements we're making as well here in the plans but we're going to try and make sure we have you know automated deployments of GitLab onto at least the two major providers right so AWS and GCP so the AWS you probably heard of with a quick start right and so we want to make sure that we deliver on that here and they continue to improve it 2018 and make sure that you know you can deploy GitLab with again like a kind of best practices with simply going through a wizard entry of the information choosing some options and hitting go and it then takes care of the rest and the same model we're trying to shoot for for GCP as well so it can help to guide you and doesn't make you go in and set up each individual components from networking to instances and things like that yourself so you can take care of that for you and stitch it all together and again reduce a lot of the initial setup costs here for GitLab and make sure that folks are you know on the right path as far as architecture as well so those are two big projects there that both for our you know for our teams will have initial offering costs as well as ongoing as you know they are using different tool sets different scripting and things like that but it does let us leverage the best of those tool sets and what they offer customers as far as features and Aurora RDS and things like that next up is you want to make sure that more of our customers can easily enable HTTPS right so make sure that they can take advantage of that for security reasons best practices but also things like the registry really really want to have SSL turned on and so you know for folks will try that you got to make sure it's as easy as possible as well and so we're working with let's encrypt hopefully so that you can simply enable let's encrypt and we'll go ahead and provision certificates automatically for you without you having to worry about going and getting them putting them in certain locations and then adding the configuration files so that's pretty exciting there to make sort of what's what's right now a little bit more challenging step much easier the next one is perhaps really interesting and this is kind of more of a still to be firmed out but we want to try and start moving towards a bit more of a kind of web admin type console right now we have you know our our github rb which contains a wealth of features and as you kind of mentioned earlier with the roles some of them interact with each other and it'd be nice if we could have sort of a simplified sort of initial configuration to get the kind of lights on and then you can then folk and prod for more of the kind of feature based settings within a kind of web admin console to do things like configuring SSL or some of the other things that don't you know necessarily you know kind of basic foundational things so that should help make github much easier to work with as well as hopefully also make it easier to sort of insert a new node plug in you know a georeplica and things like that as well here so trying to reduce some of that manual interaction manual steps required and the final one also sort of a big goal here is you want to to make it easier for customers to stay up to date on the latest version of gitlab you know if you look at some of our data we do have you know a lot of customers actually who who do stay up to date around two-thirds are within you know a handful of releases but we do have a number of folks who sort of maybe forget about updating gitlab and we want to make sure that we have a good experience for folks and really help them you know make sure they stay at the date and leverage the best new features best experience and so their developers are productive and progress they can be and also of course they get things like security fixes unless it was possible as well so we're trying to think of ways that we can perhaps do this automated in some ways at first maybe some patches or things like that or maybe you know making the ui a little easier to hit the button to upgrade for trying to think of ways to go through and do this and you can see and and join this question there as well to help increase the number of customers who are staying on a recent version of gitlab and again reduce some of the maintenance overhead of managing that process as well so some key goals there for 2018 I think they are really awesome places to spend our effort but we'll open it up for questions now as well for folks who want to comment ask some questions maybe share some concerns with you know perhaps where we're going or things that we aren't doing but that but yeah no questions we'll count the three here and if there's no questions we'll wrap it up okay thanks everyone see you in the team call and thanks mine for for as well