DevOps Unbound Special Edition from KubeCon Paris 2024 – DevOps Unbound EP 44

Hey everyone. Alan Humel. I'm back here live.

We're back here live at CubeCon. This is by far the busiest show floor I've seen in any cube cut. I know there's 13,000 people and I think they're all right here.

Yeah. Um, luckily the magic of technology, you're hearing us and seeing us, I hope. And because if you will hear the d of the background noise here is ridiculous.

I'm really happy. We're gonna be doing a little bit something special here. Right now we're gonna be doing a DevOps Unbound live kind of panel live at CubeCon.

We've got our very special VIP guest, Martin Cost, who evidently when he travels at events, brings his own security and, uh, Mihi, Mihi and, uh, hash. These are your friends, b***h. We're Guys, thanks for keeping us safe.

We appreciate It. Thank you, gentlemen. We appreciate it.

You're good guys. I don't know, I guess Martin's a high visibility kind of target or something here, huh? I Don't know what you're into, Martin, but I'm glad you're, you've got Security there.

This, like, something outta the blacklist or Something. You guys are too funny. But, um, anyway, so this is Martin Klaus from Tricentis.

If you've watched any DevOps Unbound episodes, you might have seen Martin Aren't a few of them, but, you know, it's funny. Zoom, you only see people from here up. He's actually a pretty tall guy.

Um, speaking of tall guys, to my right here is our CTO at Techstrong and Principal Research Analyst, Mitch Ashley. Mitch, thanks for being here. Always, always.

And, uh, we're here. So, DevOps Unbound, look, we always discuss everything under the sun on DevOps Unbound. But Martin, I wanted to focus today's talk a little bit on what's different about testing and continuous testing in a cloud native environment versus any other environment.

Yeah. So first of all, uh, Alan, Mitch, thanks for having me. It's great to finally meet them in person after many, many years.

Yeah, it is. Yeah. Uh, and it's unbelievable, the energy, the vibe, the community of coupon, as you mentioned, 13, I think it's actually more than 15,000 people are here.

Yeah. This place is crazy. It's unbelievable.

If you have not been to Q Con, you definitely to come here in person and see for yourself. But coming back to your question, what's different? So I've been in cloud native for a long time.

I've worked at Red Hat. 9. And, and so I would say if you look back over the last couple years, a lot of things have changed because, uh, one of their main objectives of DevOps was to, to bridge the gap between Devon Ops and to move, uh, you know, applications into production value in production much faster, to deliver more value to business, much more quickly than in a traditional waterfall or even an agile model.

And so what happens is when you're trying to move faster and you change your process to work more efficiently, to work in small increments, then the application architecture evolves as well. Right? We have moved from, you know, monolithic big web applications to microservices that you can update and scale out much more efficiently.

Data has evolved from data rest to data in motion. And streaming of events has become one way for data and applications to communicate with another. And so if you think through that, then you also have to think about how does it impact testing and test automation?

And the whole notion of quality engineering becomes a lot more important. 'cause you have to think about quality from an architecture standpoint. You have to think about quality from an end user standpoint and, and the experiences to deliver.

But you also have to think about how can you ensure quality through the development process as it relates to functional requirements and related to business requirements as it relates to performance requirements. And a lot of companies are also dealing with security and compliance and governance. So all these things have to be considered.

And, uh, it cannot be solved by tools alone. You have to think about the process as well. And, uh, most important, I would say is adopting a quality first mindset and not just, you know, move things in production and see, see if it sticks.

Um, but be more thoughtful about how you're building applications, how you deliver it, and ultimately what is the value delivering to our end users. Yeah, I I'm curious listening to your thoughts on that. You know, thinking about the people that we've interviewed and just a few of them today already, you know, it's OpenShift Red Hat folks, it's ARM processor and getting more applications on, on that platform.

Um, it, it, it's, um, you know, cloud native application companies, it's object storage, the number of variables. I mean, there's always been a large number of variables Yeah. To test for environments.

It seems like that's even bigger. I mean, you know, maybe exponentially more complex Yeah. In this cloud native environment.

'cause there's so many different platform, mean ai, you add all that to it. Yeah, I, I agree. I think if you just walk around the SHO flight here, you'll be blown away by how many different use cases and features and capabilities that are enabled on the community platform.

I think one other aspect that's really important to think about is, you know, develop or testing is everybody's responsibility or qualities of its responsibility. Not just, uh, the traditional QA folks, uh, it's developers, it's project managers, it's release engineers, it's SREs and everybody else that's involved in it. But one key difference, uh, oftentimes is, and this is something that I hear a lot from, you know, my quality engineering counterparts at, you know, not just testing companies, but the enterprise.

That it's not just about the happy path, you know, that you envision as a developer that your users go through. Uh, quality engineering is a lot about, you know, finding out what are the edge cases where things might break. What are the environments where you, you run into a gotcha type situation, and how do we prevent that from happening?

Because ultimately, uh, end users all have different environments, different browsers they log into through mobile devices, through different network configurations. Uh, they may have multiple applications running and inevitably somebody's going to use the application in Wave it was not intended or designed to deal with. And, and so that's ultimately causing quality shares.

And so that's really, I think what quality engineer to be about, is to figure out how can we maximize the footprint of all the things that we're exposing the application to, to identify those weak points that we need to prevent from happening in the first place. Martin, one of the trends that I see at cloud native con con already is the, the move to platforms, right? And I think it's a maturation phase of, you know, we're doing this, Mitch is heading up this report we're doing called DevOps next, where we're looking at what's next in, in the gamut of DevOps.

Yeah. Right to left, left to right, you know, the whole thing. And one of the trends is, Is Instead of cobbling together, we've got some, I feel like we're on the airplane, you know, I feel like I'm at the airport.

Yeah. I'm ready. I gotta go board.

See you later. You're you're missing your flight. Yeah.

But the thing I left my wallet, it's a TSA booth. That's right. Um, but one of the trends is getting away from point solutions that are cobbled together to platforms and almost like the Russian nested dolls, platforms of platforms.

So you can have a continuous testing platform like a tricentis, and that has a complete suite of tests optimized for a cloud native environment for what you're doing. And that has to fit into a larger cloud native platform of, you know, that takes up my whole CICD, my whole software supply chain type of environment. What's Tricentis doing?

Going to expand that testing platform, but also to fit into that larger cloud native? Yeah, I think the, the notion of PLATFORMIZATION is, is one of the hottest trends in industry right now. You hear about platform engineering, you hear about developer platforms, you hear about, uh, quality platforms.

And the main driver for the adoption, uh, or the growth of these platforms in history is really like, what can we do to help help developers or testers and project teams stay in a flow? Because there's so many things you have to think about. There's so many distractions.

There's so many sort of other tasks we have to do during the day. And to stay focused on a particular pool request or a particular project or a particular sort of feature requirement is it's very hard. Uh, if you think about all the meetings and disruptions that you're gonna deal with every single day.

And platforms can help with that to help you simplify and automate a lot of things so that there's less, uh, cobbling together that you have to do, uh, on, on your own. And one thing that is really important, uh, I think also from a standpoint is, uh, reusability and how can we enable more use cases with quality platforms? And, and, and one thing that we see a lot with our customers is that there are not just applications you build inhouse in the enterprise.

They're applications that, that you or SaaS applications you configure and deploy or enhance, uh, or an application you just use. So you'll see, really see the whole gamut of applications that you deploy and configure, uh, that you use across the enterprise, but then also applications that you extending and building yourself and how can you deliver an end-to-end quality automation platform that supports all different use cases between enterprise IT back office applications like Oracle, NCV, uh, business applications like Salesforce and ServiceNow and many others, as well as the customer applications that you use internally. But that can also be client facing, right?

And what we found is that a lot of customers are looking for ways to, uh, standardize process through a quality engineering framework, but they're also working to standardize tools and platforms so that enable quality engineers across many different teams to work together with their development counterparts to release applications faster. And so that's kinda like what we're trying to do to really focus on usability and the end user experience, but also make it easy for people to get started with functional automation, but then expand into, you know, data testing for example, or a load testing, mobile testing and other things, uh, that, uh, they may need to do for the particular application that they're trying to release. It really makes the case platforms aren't just for infrastructure.

Yes. No, that's part of it, but it's platforms up the stack. Mm-Hmm.

Well, it's horizontally to vertical. Exactly. It's the key to it.

And like I said, it, it's like that Russian nested dolls thing that's platforms within platforms and, but again, for Dune, you gotta make it a dune plans within plans, Spirals within spirals. Yeah. Or think about this as layers, right?

Like the, the networking layers of, you know, L one through L seven might be one analogy to think about what you doing on the infrastructure layer, what you're doing at the application layer, at the testing layer, and so forth. Well, at the end of the day, the spice must flow. Let's See.

Sleeper must awaken too, but, so can I ask the AI question? We haven't talked about ai No, it's, it's enough time. Go ahead.

We have, we have an upcoming live session on testing AI in your applications. I'm curious your thoughts, and I wish I had the, the date on it. We'll, we'll, we'll get that out to everybody.

What are your thoughts about when you incorporate ai AI into your apps? There's models, there's data, there's training, they kinda have their own flows or, you know, even more so than just a database or data source. How do you think about testing in a, in a cloud native world that has AI part of it?

Yeah, that's a great question. I think that's, let's Answer that in five seconds. The whole industry is sort of, uh, thinking about that right now.

Because on the one hand, today, for example, we saw many great use cases for AI models and for them to production and, uh, you know, using LLMs to summarize, uh, a live fixture of what was being seen. And, and so also we're seeing, you know, the use of AI from a co-generation standpoint, right? And there was a recent article from, you know, Joe Vanai who mentioned that, you know, we've shifted the problem, you know, from development to QA and what used to be, you know, three developers, one qa, it's now one developer and three QA because, uh, developers can now generate so much more code using AI tools.

And now that puts the burden on the, on the testers to verify, you know, the increase in credential security, uh, flaws and things of that nature that introduced, uh, if the code is not being tested and validated, but generated from some random source. So that's one aspect. But I think the bigger question in industry right now is how do we test the validity of the results of an AI model?

And if the summaries and, uh, all the things that an L one can generate are, uh, meaningful and they're not, uh, they're free of hallucinations and things of that nature, I think that's a great use case for exploratory testing. Um, I think there's also an opportunity to actually use other LMS to test the output, you know, from LMS and see if there's a consensus around among LMS on the results. But the end of the day, um, you know, uh, you still need a human in the loop.

Uh, you cannot, we're not at this point yet where the human is completely eliminated from the FS cycle. And as the, the technology matures, I'm sure there's gonna be more opportunity for automation to test ai. But for now i's say we're still very much in sort of the, the good old, you know, you know, exploratory testing, being able trace back to the data, you know, how the LMS came up with certain results.

But uh, you know, that is still sort like the big, the big problem for administrative crack. Makes sense. No doubt.

I mean, we can stay here and talk cloud native and AI and q well, everyone else here is testing. Actually AI was the star of the keynotes today too. Yeah.

That Saying we heard. Anyway. Hey Martin, I want to thank you for stopping by.

I can't believe you came all the way to Paris just to do a cloud native Yeah. For this one session. This is great DevOps unbound thing with us commitment.

That was nice of you. That's commitment. Yep.

But seriously, thanks to you and Chase as always. We're gonna be back to our regular schedule of DevOps Unbound, I guess, when we get back in another week or two. And, but until then, we'll be here, live all week, covering what's going on at CubeCon.

Many thanks to Martin Klaus and Chiantis Mitchell. Ashley and Alan Shimmel. You're watching Textron tv.