 Rhaid, rydw i'n mynd i'ch gael yn oed. Mae'n ddiwrnod arweinio'r cyd-defodol. Felly rydw i'r cair yn Sambrook Smith, ac yn ynglyn â ymddangos cymdeithasol cyrraedd. Rwy'n gweithio'n gweithi'r gael. Rydw i'n gweithio'n gweithi'n gweithio'n gweithi'n gweithi'n gweithi. Rydw i'n gweithio'n gweithi'n gweithi'n gweithi'n gweithi. Yn fyddi gyda'r newid o cysylltu chi yw'r platform. mae'r gweithio'r gwaith yn tynnu gwlad ar gyfer hwnnw i miwghaith y flwyddyn sydd yna ydych chi'n rydyn ni'n gweithio'r papyrwyr o'r cycry yahosu mwyaf ac sy'n dwylo'r gweithio gyda'r llewn arall yn cael ei fooch. Rwy'n gofynol i'r Llyfod Richard, Ym thyf a'r Bary, wrth i'r cyllsig, wirwch i'n mynd i ddwy'r threeo crossed'i ein rhwng sy'n oed i'n gwlad o platform. yw hwyl drwpwyr, yw hwyl heddilydd drwpwyr, yn rhan o jes, yw ddech chi'n gweithio'r pihp, yw Magento, yw gyda'r eisysig, mae'n gweithio'r bach o'r teb yn teb yn y platform. Fy enw'r ymgyrch a'r wyf, mae'n cofnod o'r eisysig a'r gweithio i ddech chi yn cyfle. Mae'r bwrdd yn cael ei hwn i'w mwneud. Gwesiaeth yna lle byddai chi eich grun o yr or estou a'r bros newsaiddol? Mae'r unrhyw oedd wedi eu bod yn rhoi eu gwneud yn y gwasanaeth, ond mae'r cael ei chyreun sydd yn cyflwyng rôl o'r cyrdeil gweithio hefyd. Mae'r teimlo maen nhw yng Nghymru eich meddwl yma. Fy ffwysig o'r technologiwn i'r gweithio, fel hynny, mae no JS, symfynu, Ruby, Python, Pats Magento. Rwy'n credu, mae'n gweithio'n gweithio'r past, felly mae'n gweithio'r fautymau arall, a'r gweithio'r gweithio. The last thing you really want to be doing is issuing tickets to your hosting binder to make configuration changes service changes to give you more environments to upscale your projects that's just the last thing you want these tickets take a long time to turn around and a lot of mistakes can be made in that manual process you're also looking for the best automation Mae'r cyfnod yn dwy ffons, ydych chi'n symlifio'r ffordd y dyfopwyr y dyfopwyr y dyfopwyr, y dyfodol y byddai'r cyfnod ar y bydd. Mae'n ddigon i'r problemau i'r ffordd yw'r cyfnod, a mae'n ddigon i'r byddai'r cyfnod ar y byddai'r cyfnod ar y byddai'r cyfnod ar y byddai'r cyfnod. ac mae'r bywch yn gwybod fod â mynd i ddylch o'r cyflaidau hynny, ac mae'n gwybod bod yn ystod am gael o'r rai gweithio. Mae'n gweithio'r ddaeth, ymlaes o'r cyfeirio, a'r ddweud o'r ddweud yma. Yn ei ffaint o'r ddweud o'r bwysig o'r ffordd, mae'n gweithio'n gweithio'n gweithio'n gweithio'n gweithio. If you can merge test, deploy in a repeatable deterministic manner and trust the results, you're going to do more of that, you're going to do it any time of the day, and you're going to do whatever the site is doing, and you're going to do that for your volume changes and for your more complex changes as well. And that's real continuous delivery. And when you're reaching the point where you're continuously delivering, you'll know that you've streamlined some of those really complex business processes. Not only that, but you're giving your service management, your product owners, your head of e-commerce, the head of marketing, you're giving those guys the green light to think on their feet, to react to the marketplace, to make changes to the site whenever they need to. And that shouldn't have happened. Let me put that back into presentation mode. And basically, that's where you're beginning to see real competitive advantage. So just one last slide. When you're next assessing your hosting pass vendor, you need to be asking them all these questions. It's shown me the real results from all your customers, not just one or two references. How much faster is my development going to get? How much more productive will my development team become using your pass? How much better will deployment get? Will my deployment velocity improve significantly? And if so, exactly how much better is it going to get? How much better is my service going to become? And are my costs going to change? And you need to tell me exactly what my overall ROI is going to reduce to using platform or whatever that pass is that you're considering. It's all about proving it, asking the vendor to prove it to you. And on that note, I'm going to hand over to Myro. He's going to introduce himself. Myro is the founder of ND Systems. So, hello. Does the audio work for you? Yes, I'm founder of ND Systems. It's a company from 2000, started in 2008. With Drupal in 2008. Actually, we started in 2012 already. So we're one of the most contributing companies to the Drupal 8 lifecycle. We're on position two of the global contributors since, I guess, three years. And most importantly, we're living this idea of making contribution part of every project we're implementing. And so this reducing complexity from customer projects, lowering total cost of ownership, and creating a healthy environment for the Drupal ecosystem to survive and to develop, to innovate in the long term. And we were actually seeking, when we started thinking about running Drupal 8 early, a platform that fits our expectations. And we were heavily involved into the questions of how can we achieve maximum performance with Drupal 8 and without paying it too high price. And it turned out that when thinking of the capabilities of Drupal to make it super-performance, we wanted to have additional services like Redis or many cases like you wanted to have a search API backed by Apache Solar. And still, we did not want to tell our customers that suddenly their hosting is like the price of three or five times the amount of before. And that's why we basically were super happy to find the platform as such as a service provider that allowed us to deliver something with this added complexity at a similar price to what we delivered before. And also this kind of delivery basically, we didn't need to solve the problem of maintaining the services or like this. Basically we have no additional complexity to maintain these dependencies on our side. And similarly, for us it was very important that we can start with a small site, a small customer. And like we developed the first distribution, it was called NPH for news portals. And we had like super small sites that wanted to run it. And we had super large enterprise like clusters that we're running. And we wanted to have consistency to basically maintain both with say Drupal is a perfect fit with like plugging in, caching backends and stuff like this. But we wanted to maintain consistency with that distribution and basically not adopt too much within like whether it's standard or enterprise. And that's the perfect fit for us like we are actually maintaining this distribution this way. We have these super small sites and it's just a little bit of configuration differences you have consistency within these environments. And sometimes, yes, even like media agencies, they decide that we need to deploy on traditional infrastructure, but still it's helpful for us to decide to standardize the development environment. So basically all of these customers that run their own seller hosted. Yes, classic environment. We run still a platform development environment because it's easy for us to quickly fork, deliver, test something and basically have independent review cycles. And afterwards, there's still this old school process to go to production pity, but it helps us to not basically maintain this problem during process of feature development. And also here the thing is basically our customers have access to the admin UI, so it helps for them to experience like to see there's something going on, there are updates, there's activity from the developers that's really helpful in HL processes. Sometimes even customers, they fork on their own. Like their product manager, I want to test something and I don't want to break production. So like they're doing this even service. And yes, as I said before, like when we then switched from using Drupal to build highly custom things into more standardized environments and for this basically we can ship our distribution no matter what its production like enterprise or small standard things. We basically ship the same code base and it helps us to have a standardized process. No hassle custom like custom setups figuring out how to set up or how to configure to in that specific target environment. Actually this leads to very low maintenance cost in shipping out new releases of our distribution for all the customers that are built on top of it. And the next thing was like still there's this distribution, we can ship it, we can create the environment, we can initialize it, but we can go even that far and basically release such a distribution so that a customer can create an own instance as a self service. Thunder demo, like the idea was, you know, if you want to test drive thunder, what are you doing? Like do you set up a local deaf environment? Like if in case you're a product manager you're just evaluating a potential news platform. So we said we want to simplify this process. Thunder demo is about just filling out the form and then you're entering a checkout process on platform message that is basically triggering a complete install. You're owning your own environment, you can promote it, you can keep it for free 30 days, and afterwards you can decide to keep it, promote it to standard, release it for sure. Thunder is still not a product that is like you will not go live the day after. You'll need theming, so it's not out of the box perfect themed product. But at least we removed the barrier to test it and we allowed them to keep it easily. And we added some special tweak for speed up of the installation from a database dump instead of the regular installation profile. That would take too long, specific tweaks here. And then like beyond that for us, we are using bots. For instance we have like a bot address. We're just adding it as a user and if you add it, it starts as a daily snapshot or something. PT platform doesn't offer it out of the box, but at least it's an interesting thing. Through the API you can discover the project and you can implement bots that are using the API to do any kind of automation. It could be the rush export, like a backup of the system, whatever you want to do. It's really helpful. For the population of new projects, for instance we add our standardized accounts of our lead developers to all the platforms. We do most of the setup of new projects kind of scripted automated. And for sure we use the same to iterate over all the instances for like thunder demo or all these things to deploy updates in an automated way. Right. Hi there. Oops, happened again, Ciaran. Okay. Hi. My name is Richard Jones. I'm the CTI of Invika. Invika, what's on these slides, here we go. We're a technical partner agency. We work with quite large organizations. You may know me if you've known me already. We were ICOS previously before we joined Invika. So we've been part of the Drupal community for many, many years. Now we're part of a bigger organization. There's 200 of us across the UK and Germany. We primarily work in PHP, Symphony, Drupal of course, but also in other things like Magento, Spryker and Easy. So when I spoke to Ciaran about participating in a panel today, it was very much about what were you doing before, what are you doing now? So I just want to talk to you a little bit about what we were trying as ICOS and as Invika when we joined. So you can imagine we had two tech teams, one coming from a pure Drupal background, one coming from a multiple PHP background. So we were doing all sorts of things. The ICOS team were very, very into Drupal VM, which is a really great solution. So we had lots of vagrant VMs for different types of projects and that sort of thing. The Invika team had a slightly different approach. They managed their own vagrant machines and VMs because they had much more ops capability than we did. And they had their own processes that they developed for building new versions of sites from database snapshots and this sort of thing. So we had some managed hosting. We had some client self-hosting. We had some stuff with Acwea and we had some DOI, local development kind of stuff. So we were very much all over the place. No real consistency of approach and obviously when you're merging companies together, it's inevitable. You'll be very lucky if you ended up merging with a company that happened to do the same things as you. So as part of my role as the chair of the group, what I needed to do is try and bring consistency. The way we operate with this many engineers is that we often have to bring, we have to move people around. We have to move people between teams, whether it be covering holiday or if a team is ramping up, we might need to swap someone in. And the most important thing for us is consistency. We can't have someone spending two days getting themselves set up. So we're really looking for consistency of approach, making sure that you can switch from one project to another and know exactly what you're doing. You don't have to learn something new. Of course you're going to have to learn something new if you're switching between technology and we don't ask people to do that very often. You certainly don't switch from Drupal to Magento overnight. So we do have specialist engineers but really what we're trying to do is improve that on boarding time. Making sure that we can get a new engineer, either a new starter or someone moving between projects as quickly as we possibly could. Traditionally you have this sort of dev stage live set up, whether that be through one of the other providers or as the clients provided for us. Sometimes dev might be local, hopefully you have some sort of staging and UAT environment available to you and of course production. But the way we wanted to work and the way we do work is very much about true agile development where you are trying to get each individual feature out there into QA. So we needed a way to make that possible so that we can actually have basically our own QA team plus the client and the product owner looking at specific features in development. What was happening to us before is everyone's competing for the staging environment and everyone's getting confused, banging into each other and no one's quite sure what's going on. So you end up in this scenario where you only refresh the staging environment once a week and that's not really fast enough. So we needed isolated QA for our teams and we also needed representative data. So we needed to make sure that what we were seeing in QA was actually real, sanitised obviously, but making sure that it's real data that's useful and testable. So that was my sort of analysis of what we actually needed to achieve. So we've known Platform since the very beginning and we brought Vivian Westford on board very early. However, recently we've been working with some large organisations so you may have seen that we launched the Arsenal website very recently. That was on Platform and we've also very recently launched all of the gap international sites around the world and I'll show you a little bit more detail of that. So these are sort of enterprise level sites that we were able to put out there and you can imagine something like Arsenal has very peaky traffic, especially on match days. So it's very important for us to work with a partner-like platform to be able to get this thing up and stable and ready to cope with those peaks. So where we've ended up with today is we've now got a developer workflow that is exactly what we wanted. So we've got a developer workflow where we can make a new branch. At the moment someone does a pull request in GitHub, then it immediately builds the environment. Then we get posted a notification in Slack that that's happened. The QA or the devs can do a peer review of that, look at it properly and truly test it with real data. So that's where we've got to and that's exactly what we were looking for. We've also managed to plum in all of our automated testing like Scrutinizer and these other tools. They seem to vary, but Scrutinizer we have at the moment and we're able to use the hooks within GitHub to deploy all of these things. The other thing that was really important because we are not a peer-drupal agency anymore, we are using Easy Platform which is another PHP CMS. We are using Magento, we are using Spryker, but by having platform allows us to actually have a consistency of that experience. We don't have to do Magento stuff over here and easy stuff over here. So that was a real benefit to us and probably wasn't one we anticipated at the beginning, but as platform has introduced these new partnerships it's been a really good positive for us. We've also been able to do some quite complex stuff. We've got Node.js stuff working with Gap, which I think is the area it is. So this is just an example of something I wanted to show you. What we've done with Gap is this is not gap.com, this is not like the prime site. This is all of the other stuff. Gap have a model, a partner model around the world, so they have multiple sites. In this case we're dealing with at 25 sites, but they have a centralized content model. So what we've done is on platform SH we've built the content hub in the middle. So the marketing team in the UK build all the content in each season and then in the middle of the ring there, I'm sorry it's too small to read, but you can go out there and we can deploy content out to the different UAT and feature branches and staging as I was just mentioning, including training. And then the final part of the puzzle, we were able to put Node.js in place, which what it basically doing is interrogating the content hub site and selectively publishing out to all of these different country sites. Now, the reason we use the Node.js in this case is because we have to be a bit clever about what content goes to what site. Sometimes some of the imagery is not appropriate for certain regions, but more importantly, sometimes the season is the other way up. So we don't want to push summer content out to a winter season country at this point in time. So this is the type of stuff we've been able to build the platform. We are also able to give each of these countries their own individual platform account and login. And that means that they've got control. They've got control of their own hosting as a partner, but we've got control with gaps centrally making sure it's up to date and managed. And we were able to use things like the, there's a command line multi tool that allows us to push commands out simultaneously to all these sites. That's been a real, a real godsend to us. So where do we go from here? Well, where we are right now, we always try and present platform as an option because for us again this consistency, I would much rather be doing this than building something with AWS or something like that. We do have DevOps in-house, but I don't have enough of them to be maintaining huge arrays of custom hosting. So for us, this is the right balance between, as I put at the bottom there, platform SH is slightly opinionated, but not completely restrictive. So we've got some options as how we set things up, but there's not like infinite options. So I actually quite like that because it means that I know I can go from one project to another and the things we're deploying are going to work. So we know we can put something completely custom and we can start from a clean sheet and we can build them whatever they like. But for me it's about helping our DevOps and our engineers internally to say that's not necessarily the best approach for the client. Where's the budget best invested? For us, the budget is best invested in building the unique features, building what we call the business value and not so much on spending time tinkering with hosting. But that's just me. So sometimes custom solutions are required. We have looked at Kubernetes and I think that's probably a conversation from another day. Sometimes things like Site Factory from Acria are the right thing to do. But most of the time we find that platform is in that sweet spot. It's basically what we need and it gives us that value in using a standard platform for the types of things that we do. That's me. If you've got any questions, feel free to ask afterwards. Did you get anything? Actually one thing Ciaran asked me to do was to come up with some solid numbers which I haven't done. I haven't got anything that says we've reduced that dev time by 80%. But I did speak to Steve just now who is the project lead on Arsenal. And the one thing he said to me is it's okay we deploy on Fridays now. So maybe that's good enough. Okay, that's great. So I'd like to introduce Barry now from Real Life Digital. Thanks Ciaran. We have numbers somewhere in here. So hopefully mirror similar to what you've got. So I'm Barry Fisher from Real Life Digital, East Hampshire in the UK. Our client base ranges from small to medium sized manufacturing sites up to blue chip companies. One of our clients that keeps us quite busy is M&S Food. We've been doing mostly Drupal development for the past 10 years. But we're starting to look at other different platforms, so Node.js using Vue, using front-end components and different APIs and things. So I'll get on to those in a bit. But I thought what might be useful is to just highlight where we were coming from before we started using platform SH. And it's important to emphasise that we don't have a dedicated DevOps role because we're quite a small team. So it essentially landed in my lap usually when we needed to fire up servers. And that would be done in a fairly haphazard way. Very much DIY. We might have one client that was on a rack space cloud. We have another one that was using AWS that we didn't manage directly. We would have someone digital ocean. We would just be all over the place. And shared hosting, if we're going back far enough, but obviously nobody uses shared hosting for any serious sites anymore. So VPS, we're trying to get as much as we can onto platform now for the reasons that Richard mentioned, in that we don't like the variance between the different projects and the fact that if you put a project down for a couple of months or so and you come back to it and you need to make a tweak based on what the client's asking for, you don't want to sort of have to think, well, how has that done again and how do we have things configured? So previously we were using Ansible. That's where we got up to before we started using PlatformSH to try to automate some of those processes because it would take us, it would still take us half a day to a day with the Ansible scripts in place just to make sure things were working. And then at the end of the day you would step back and just hope that everything had executed in the same way that it happened previously. So that confidence wasn't quite there even with having Ansible, which should have been automating those things for us. And then we had, say, members that would come onto the team and then we would need to manage the SSH keys and secrets management. Do we give them password for this environment or how does it work? That's become a lot simpler now. As part of the development work itself, the whole works on my machine thing became a bit of a problem because we had developers that were working on Mac. One was working on Windows at the time and the rest of us were using combination of Mac and Linux. And that left a lot of sort of what we call meta work, sort of not actually providing value to the client. We were doing things that needed to be done for servers, for all of the things that the client doesn't see, needs but doesn't appreciate. So we were also finding the retainers that we have for clients were being used up for this stuff that they never see. So then questions are essentially raised about, well, where has that four hours gone this month? And we've essentially reduced that to zero now for using platform. So the result of all of this is that clients were paying for things that they don't understand or care about. So we looked at other pass services. You can see them all down in the foyer. There's plenty to choose from. But I feel that platform is sort of ahead of the curve in terms of breaking the mould with the dev stage prod approach. Because essentially if you've got a new ticket or if you've got a new piece of work that needs to be done, you just get branch your code and then essentially a new environment can be created really easily. And we've reduced that down from, if we needed to create a new environment on a VPS, typically that would then be something that comes to me. So if you need to create it, check the massage, the ansible scripts, and you could be talking a few hours just to make sure that's all essentially ready in there. So it was quite a heavy weight process. So it was an expensive branching process. So now it's just essentially git branch. That's created in a matter of a couple of minutes and then essentially we can then UAT. The clients can go on to that environment and check to see what's happening in there and we can then merge it back in. For projects we still have a staging branch but they're just easy to spin up and set down and synchronize the data between the different environments and making sure we get the latest live data and sync to cross is usually quite an important thing when we're going through the tickets that we're working on. Another situation that some of us may have come across is that if you're going through the client's IT department, certainly if you're working with blue chip companies, raising a ticket to say, let's install Redis or if we need to install or just make a small configuration change, that could essentially be weeks and that's not an exaggeration. We could be waiting really long time just to make trivial changes. So we got to the point where we looked at some of the other providers and it just wasn't feeling like it was giving us the same degree of control that we would have because obviously going with a pass service you give up a certain degree of control about what you can customize, what binaries that you can install and how you can go about configuring things, how you're used to having them in a dedicated space. So that's something that essentially made me quite nervous sort of moving to and investing quite heavily in just one particular platform was not having that degree of control and I'm glad to say that we haven't come across anything that we haven't been able to do certainly with the help that we've had from platform SH. So what we actually needed was a platform to rule them all. We needed the flexibility, we needed to be able to install things as we needed to, we needed reliability, testability, security, all of the things that end in TY essentially. And we were starting to think well what about the other technology that's out there as well? So we were starting to look at Node.js, view and view components and how we might be able to make the most of what the front end JavaScript world is providing us that can provide us with. So platform provides support for MongoDB, elastic search, solar, all things that we're either using today or essentially key parts of the stack that we're looking at using in the near future certainly in the next six to twelve months. I think that pretty much sums up what I've got on there essentially. So some providers that you'll find that will force you to use certain caching back end. So we'd already started for this particular blue chip company. We'd already built a lot of our services around Redis, which is a caching mechanism which just comes out of the box. And in terms of setting up Redis on platform, I say if those that are not familiar with it, it's essentially just editing a config file, putting a line into that essentially says I would like Redis and then Git deploy or Git push, sorry. And away it goes, it rebuilds and then it's then available to your application. And then it just comes down to a case of enabling the relevant modules in Drupal. So it really is easy just to fire up those things and not really give it too much thought. You can just consume those services that you need. We're also looking at setting environment variables which is quite useful for you're doing certain things like indicators on your website to indicate whether you're on the dev version, the stage version. Essentially you want a big red banner to say you're on the live version because quite often as developers we can have several tabs open and you want to make sure that there's a strong visual indicator to say you're actually on the live site here. So we needed the moon on the stick and it was quite a big ask so we went back to the old ways of doing things until we came across platform. So our introduction to platform SH was, I can't remember exactly how we came across it, but I decided to give it a go because it was getting quite painful going through all the dev ops. I might be able to do this. Can I? It's there, is it? That one. So one thing I will say is really quick to get on board with platform in that the documentation is extremely good. The majority of getting up and running with platform is understanding the configuration file. I don't have an example to show you, but on the documentation it's declarative in that you say this is what my environment looks like, this is the PHP version I want to use, this is the version of Redis I want to use, I'd like to use my SQL database and there are several github examples that you can go and grab so it's very easy to get up and running. I spent about a day to familiarise myself with what the quirks were, what those specific opinionated isms are for platform and the documentation gets you very, very far, it looks at 90% of the way and then if you've got any extra questions I've found the support to be very responsive. Another thing that just pops up now then on the blog post is something which I'll just mention Let's Encrypt here which is essentially a free SSL certificate service. I saw a blog post come up that says we're now offering this as a matter of course and you can just put in your routing file that I want to use HTTPS and then your SSL certificate just appears out of nowhere, it just works, it gets renewed, it's free so it's just a case of declare I want to secure my site by putting one character an S in and everything seems to work. I like the fact that the fact that when you push it's essentially a fresh build so that would be taking your repository doing a composer install doing any scripts that you need and again you just declare what you want to happen in terms of a build and what you want to happen on deploy and there's a slight variance in those two so I like the fact that it rebuilds everything and it forces you to use best practices in that the read-only file system that you get for the code so that is essential for vulnerabilities that come out like Dribblegeddon is that the code base is essentially not writable, you declare where your writable areas are and where your non writable areas are when we're working with teams we may have a developer come on and then leave again so we just use their public SSH key again very easy just to integrate that and we still need to use the tools that we would use previously which is the rsync tools which synchronises your files between what's on your live server and your local and SQL sync which synchronises your databases but a lot of that the drush aliases that come with the platform essentially set all that up for you and then you can just pull down so that you make sure on your local copy an up-to-date copy of the site that you're working on or the issue that you're working on so the outcomes of that were there was very little onboarding support that we needed and we felt confident in how easy it was to get going the result of this was we were spending less time doing technical or server maintenance tasks so we were happy, our clients were happy because essentially we've saved ourselves time and we've saved our client money from their monthly retainer because the way that our business model is is that we have a certain amount of hours a month allocated to a client and that allows us to do more to help them so it may be that we've got extra features that we might be able to bundle in that munch or book in another meeting with them so that we can help them move forward and this helps us stay competitive as well because our competition is still wrangling with these issues so in terms of numbers I look back at these I thought they look over the top but let's see if we can break them down a little bit so 80% reduction in overall total cost of ownership even for a basic site so like I mentioned about the client support is the amount of time that we wasted on doing DevOps is that we've managed to get a day server set up down to less than an hour 8% 8 times faster to set up new projects there we go that's the day down to one hour and once you've got one platform site up and running with a config file you can then just essentially copy and paste that tweak to your heart's content and then you can just use that for your new project it's that easy 6 times faster deployment time usually deployments would come via we would book them in and we wouldn't deploy on Fridays and that whole thing so spoke to the developers that we were using on a particular project and we said well how much quicker is it using platform SH versus what we were doing before and they said well they are actually saving half the amount of time they said half the amount of time that they were spending were wrangling with merges with code that wasn't working with other branches and so on but because branches are so cheap and easy to create and throw away now we're producing new branches a lot quicker and being able to merge those things in and this is the one I had to think about because 60 times faster set up for new dev environment so we have got that process down from half a day down to four minutes so it took us half a day before because again it would come to me I'd have to create a user create the set up the account make sure it's working and essentially now it's just essentially git branch and you've got a new environment and you can get up and running it's not it wasn't all the time that we were creating environments because it was taking this long whereas now as part of our process we just embrace that, git branch as much as we like create as many environments as we like and then merge them into a single UAT branch and then we're happy days so where do we go from here so Drupal is the number one CMS and has been for a while but clients are asking for different types of solutions so they're asking for chat bots offline tools if they're for example one of our clients M&S food, they have auditors that go on site to different factories and they want to have a version of the audit form on a tablet offline because a lot of them are in remote environments and they're all around the world they want to be able to fill out that audit so we need to look at local storage and then how do we synchronise that when that's online and that gets outside of Drupal's comfort zone so we're starting to look at how we might be able to create Drupal and Nuxt which is a view based framework which is still fairly new but we know that by using platform that supports Node that all of these technologies are essentially they've been stabilised so that when we come to it any variances will be just down to what we're doing with it rather than the actual platform besides we haven't actually deployed any Node.js apps so we'd have to look into how we would actually do that in the traditional sense when it's just up and running so that's one less thing to worry about I can mention with the blog post with Let's Encrypt that it just works there's a few little things that other services that I've found don't include so that might be image optimisers and the small binaries that you might use with some edge case Drupal modules that you can bundle up and use your own local binaries so long as they're in a particular format and that can go up with your code build scripts and then you've essentially got endless options that you can use custom PHP extensions which aren't part of the core of PHP you can use custom binaries one example is the search solar API attachments which will essentially take a PDF file unwrap the contents of that PDF file and put it in a format so that it sits in your Apache solar index so it's a bit of an edge case but still a lot of people need that and we can get that working with no problems at all and something else that's on our radar is to investigate how we can run our automated tests as part of a development workflow because Platform doesn't provide that as part of its service but it does integrate with webhooks and the API is extremely rich in fact it's an API first so everything you see in the UI is available in the command line API and gracefully we're at the end right great that's great thanks very much well thanks to our three presenters I think we've got time for questions and we also from our three agencies but also we've got Rob Douglas in the room I think he was in the room isn't he no he's left anyway that's a shame he knows an awful lot about our customers and products and what we're doing next before we ask any questions can I have a show of hands who's familiar with Platform and using Platform right okay this is a fair few that aren't and those who are now interested in using Platform can we see some hands okay there's too many hands that didn't go up in answer to that question so do we have any questions Platform or our presenters before we wrap up yes we have the sorts of ROI that sort of ROI that Barry put up we actually collect metrics in ten categories we call it data driven hosting decisions we collect metrics in ten categories overlapping those and we probably have about 40 customers that have put quite a lot of thoughts into that they've been using Platform for several months in production and we've actually we're doing case studies where we've asked them to think about the productivity improvements the savings deployment improvement, velocity and so on and so on so we do have that and we're building that out fast any other questions okay well thanks for listening I think we're done thank you very much