The Lab with Brendan O’Leary EP 18

Hello friends and welcome back to the lab a monthly show where we look at the latest and greatest developments in software development devops and the cloud I'm your host Brendan O'Leary coming today from a very wet Annapolis, Maryland. It's been raining. I think for the past three or four days.

And for those of you who don't know me. I'm a Staff developer evangelist at gitlab which means I get to talk to amazing folks throughout the software engineering space about what it really means to get code shipped into production. 4 releases out and you know brings a lot of new cutting-edge devops capabilities to get lab and to the gitlab one devops platform.

And so we've got exciting product updates that's going to help your developers and your teams collaborate and deliver software more securely. And I think that's what we'll get started with. 4 comes that our first ever machine learning power feature improve ci/cd integration various security and workflow features and a lot more.

But to talk through these new updates. I have the pleasure of having with me today Derek Ferguson senior product manager at gitlab who's join us to discuss some of gitlab's latest features and what's to come from gitlabs. So, let's Dive Right into that and kick it off with the big one.

As I said this releases the first from gitlab to offer machine learning power machine learning power feature. And we know from gitlabs 2022 Global devsecop survey that Ai and ml use is on the rise and software development. With 31% nearly a third of teams saying they were using AIML for code review, which is almost twice as many as said so last year in our survey.

And so Derek as one of the product thought leaders in this area. Could you you know, tell us a little bit more about this this new feature. Yeah, definitely am happy to share more Brendan.

Thanks a lot for the invite. So yeah with 15 four gitlab can now recommend merge request reviewers with a feature that we're calling suggested reviewers, which is in open Beta currently. com customers.

And just a little background. I guess the suggested reviewers is the result of successful integration of the unreview acquisition. So we acquired unreview back in 2021 and it gave gitlab practical way to get started with machine learning with a really focused value proposition.

That was pretty obvious to the platform. I think. Yeah, that's it's really exciting to hear.

I was really excited when we had announced that acquisition and even more excited to see it come to fruition. And in fact, like a little inside baseball is we've been using this feature ourselves, right the beauty of building gitlab with gitlab when you work at gitlab. Is that you get to use some of these features before you know, they're out because we use gitlab ourselves to build gitlab.

And so this suggested reverse features something that I've seen for a while. I'm really excited that now we're we're releasing it in this in this public beta Dare, could you also tell us like how that came to be and like what's next for you know AIML for forget lab? Yeah, absolutely.

Yeah for devops teams. I think that you know anybody that's that's worked within that realm can. Say that, you know deciding that the right person to review a merger Quest isn't isn't really always straightforward or obvious.

I mean merge requests within gitlab are the primary method of making changes to files in a project. So making merger view or the merge process, I guess more simple and seamless something that we're always iterating on and improving. So like I said, you know choosing a reviewer isn't always straightforward, but it's definitely not something that you need that you can ignore or that you should do in haste.

I mean, the wrong reviewer can cause delays or low quality reviews back and forth reassigning reviewers and all that or you know, you can even end up with no reviewers at all. Which is definitely not a situation you want to be in so by automatically recommending reviewers, we can remove some of those menial tests that get in the way of Developers workflows. And we're super excited to continue expanding gitlabs automated and AI ml-powered features and future releases.

There's definitely a lot of demand in this realm for software testing. For example, 37% of devops teams use AI ML and 39% plan to introduce AML powered testing in the near future. So there's definitely an appetite for AIML and these features that create seamless software development experiences.

So we'll definitely keep delivering that with our one devops platform. That's great. That's great, Derek.

Thanks. And you know normally on the lab I'm talking about kind of my favorite features, but maybe you could let us know, you know, besides this. What are your favorite a few of your favorite features from the latest release?

There. Yeah. Um, yeah, I can tell you about a few of mine.

So one of them is that you can limit the maximum number of custom domains per projects. So you've been able for a while to use gitlab pages to Define custom domains for your websites, but you know too many custom domains can historically or has historically resulted in slow response response times for the pages API and you know, it's impacted the overall reliability of the service. So now you can limit the maximum number of custom domains per project at the instance level so you can strike the right balance for your needs.

The default is zero, which is unlimited so you can you know change that whatever whatever you need. And then another one I think would be getting started with gitlab pages. So it's it's a whole lot easier now with 15-4.

And instead of creating the configuration files by hand, you can build them interactively using the gitlab UI so you just have to answer a few basic questions on how your app is built and and we'll build the gitlab CI yaml file to get you started. And this is really the first time we've been using our new pipeline wizard. So it's a tool that makes it easy to create gitlab ciaml files by building them directly in the gitlab UI as you can definitely look forward to more more simplified onboarding helpers like this one in the near future.

And then I guess the the last one would be that we've improved the ci/cd integration in Visual Studio code. So when you're building up complicated gitlab CI yaml configurations that you know might include something like include or extends keywords it it can be challenging to ensure that the configuration is valid and that the file has your expected configuration. So you can now use get lab workflow for vs code and preview your gitlab ci/cd configuration file directly in vs code and you can even view your changes locally to ensure that you're configuration is what you expect before you commit and push And there's even some more gitlab workflow updates that provide more cicd pipeline interactions to help you avoid context switching and things like that.

So I definitely recommend that listeners check those out. Great. Yeah, thanks for that insight.

There it Derek and I think it's really interesting those last two you shared, you know, the gitlab c i yaml files so powerful, right but it can a it could be intimidating for folks getting started. So I'm really excited again more more now. I'm stuck on this baseball too inside baseball here.

Is that the the pipeline whether it came out of our incubation engineering department. And so I was I've been following that really closely as well really excited to see that come so that you don't just, you know, a lot of times we talk about in anything any kind of you know, knowledge work starting with a blank sheet of paper can be the hardest and so just this little wizard to help you get started when you have a blank cimal and you're just trying to figure it out. I think that's a really powerful feature that I think is gonna really help teams adopt right more of the one devops platform a lot easier over time and then kind of on the other side of that right when you're really Advanced devops team or you have a very, you know, complex cicd integration.

Anyone that's ever looked at gitlab the project can see there's a very complex CI setup there, you know giving tools for those folks to be able to manage their cicd configuration even easier. It's kind of like both ends of the same both sides of the same coin there. So I'm really excited about all those as well you hit it right on the nail on the head and ci/cd is always really near and dear to my heart.

So that's great. And thanks so much for joining us Derek. I really appreciate it.

We'll have you back anytime you'll you'll have us and there are a few more features from this month's release that I do want to hit on for our listeners. Although I didn't really like Derek. Um one is the ability to sort and filter tables that have data in them in markdown.

So working with table. I love markdown but working with tables and markdown like to call it cumbersome is is maybe an understatement. It can be difficult to figure out.

You know, how many pipes you need and how many empty cells there are but then that table output is also very static when it's you know out in whatever document if it's an issue or the wiki or somewhere else, but now you can easily insert data-driven tables using a Json syntax that lets you, you know save issues submit commits or publish the page and then you can sort or filter on those fields. And so it's a really simple you could just click and have the resource, you know, 100 rows in a table just as easy as any, you know web searcher that kind of way we're used to interacting with the web but with markdown, so I think this is a really great feature again bringing those kind of two sides of the same coin Simplicity marked down, but then the, you know complexity that you can have when you have more data together. You also can now have that users on a verified domain can bypass the email validation.

So This is really helpful. Like if you get a new users created through saml or for through, you know, some sort of other centralized provider that belong to a verified domain that you say, you know, this domain is safe. They don't have to also get a gitlab verification email which just reduces the friction right accounts that are generated through some sort of provisioning process like new users onboarding or bringing a new team in to get lab.

Those are already verified. So users shouldn't have to individually also verify them manually, so I'm really excited about that. You've also can now add linked resources to incident issues.

So incident issues to remind you are specific kind of issue involved in The Incident Management and features of gitlab. And with this release you can more easily surface, you know, the incident slack channel the zoom meeting space or you know links to a specific resource for resolving instant instant instance. And that's really key because you know being able to collaborate and be able to be efficient in the middle of an incident are really critical.

Critical path items to take care of and users don't want to spend time, you know setting up a new collaboration tool for each incident and you know, so this feature really helpful helps handle those roadblocks and unblock users to actually focus on the incident at hand. And then last but certainly not least again ci/cd and and get lab Runner Fleet is always near and dear to my heart and I'm really excited that there are now more powerful Linux machine types available for gitlabs sass Runners. com or get lab software as a service.

You know, you may see that you've got these shared Runners that get lab provides where you can run your jobs on but those have kind of always been just a homogeneous set of machines, but we now have access to you know, medium and large more powerful machines so that you have more choices about how you want to distribute your CSD jobs. And of course, you still have the great, you know, 100% job isolation, everything's in ephemeral virtual machine. So, you know, the security and auto scaling is all automatically handled fully by get lab, but you also get the benefit of of you know more powerful machine types.

So it's again Best of Both Worlds situation there. So that's a lot of the great features that were in the latest release, but of course we can never get through all of them in our short time together. 4 release blog for the full list of over 60 plus updates.

And Derek again. Thanks again for joining. We can't wait to have you back on the show sometime and thanks to all you viewers for watching today.

Again, you can find me on the internet. I'm at O'Leary crew most places and Twitter's where I spend most of my time. So if you have an idea for a future episode or want to discuss anything from today's episode in more detail, you can find me there my DMs are wide open.

Again, the lab is produced monthly. So thank you again for joining us and I'll see you next month in the lab. Stay safe.