Harnessing Innovation: Insights from KubeCon Day One | KubeCon SLC 2024

This is Textron tv. Hey everyone. We're back here live at Cube Con for our continuing our day one coverage.

What a day one it's been. We've had so many great people. This is Cube Con, right?

There's always an energy at this show. You guys have been to Cube Con before. It is.

Uh, and I do a lot of conferences and it's fun going to conferences. But this cube con, no matter whether it's in Europe or here, there's always a, a hustle and bustle energy that you could feel. And this certainly here, this way this year.

Let me introduce you to our two guests. I'm gonna start on my far left. Hemanchu.

Hanman. Yeah, Han. Okay.

And this gentleman here is Yuri Shine. I get it. All right.

Both of them are with Harness who always loved working with Harness. I was telling Uri I remember. com 2013, we launched the first time we published March, 2014.

We had two sponsors on the, on the site. First sponsors, one was AppDynamics Oh, brilliant. From Geo d Bonsai.

And eventually that of course got sold to Cisco. And Jody was the founder of Harness. So I've been following Harness since the day it launched.

Right. And I remember, I still remember my interview with Jody. He felt like cd we CD had had become people just thought of CI ICD.

You didn't have CD without ci, but you had CI without cd and he just thought there was a better way. There had to be a better way. And that's what Harness was going to explore.

He was right. Yep. As, as Jody often is, Jody's not a dumb fellow, right?

He's a pretty smart guy. Correct. Um, but as we sit here now, a lot has happened, right?

Cloud native, right? The whole rise of GI Ops and Kubernetes and you know, the whole way we publish software, we deploy software that's changed so that it obviously has changed. CICD software, supply chain, all of that.

But let me not get ahead of myself 'cause I want you to talk to harness story, not me. But first let's talk a little bit about you, URI. Let's start with you.

Tell us a little, your background. Yeah, sure. Absolutely.

Thank you so much, Ellen. So my name is ui, uh, VP of Product Management at Harness. I joined Harness, uh, about five years ago.

We only had one product back in the day, which was, as you said, continuous delivery. Yep. This is the first product that Harness launched and since then we grew our portfolio quite significantly.

We now have a 15 different products Wow. All under the same platform. Yeah.

So very comprehensive DevOps platform. And, and I, my back, my background with DevOps, uh, start, uh, way, way back, I think I've been doing DevOps or release automation as it used to call about. I'm sorry we used to call it Yes, yes.

Back in the day before Kubernetes, before containers, before cloud native applications, uh, I was doing that and now I'm excited to be at the cutting edge of technology, which is what Harness is doing around continuous delivery and DevOps in general. So that's very good. That's Me.

Thank you. Sohu, what about you? Nice to meet you.

So yeah, I'm Iman Chu. I'm a staff direct manager at Harness. I joined on about two and a half years ago.

Um, I was an engineer back at Spotify and I was, um, um, thinking about how the whole cloud native shift has shifted all the technologies so left that the developers are not overwhelmed anytime they have to ship a feature work on something. So there has to be some sort of abstraction between what developers want to do and where, what the platform engine wants to do. And I saw that at Spotify.

It was amazing. It was super easy to create a feature, create new software, but most of the companies studio are not there yet. So they had something called Backstage, um, and, and I was an engine on that.

And in 2020 we opened Source Backstage and it became a CNC of Incubated project. So, uh, just yesterday I gave a talk on Backstage adoption. Um, and yeah, that, that's how I got into Harness.

Then Harness decided to take Backstage Project and ensure that our customers are able to reach that vision without having to have 15 engineers like Spotify has, or, you know, dozens of engineers work on their stuff. So it should be easy to reach that vision with Backstage and have an amazing developer portal and have effective developers. I love it.

So you, you mentioned 15 products, but you you's one Platform. Correct. And all 15 of these products can actually work within that platform.

You may not need all 15. Some people, you know, I always used to have a, a discussion. I come from the security world.

I spent 25 years in before it was called Cyber InfoSec. We used to always have the discussion feature versus product is something that, you know, what's, today's product becomes tomorrow's feature, right? Yes.

When we look at these 15 harness products, are they really features of the, of the platform or do they stand alone? Yeah, no, that, that's a great question. So actually all of these products are actual products.

You can consume them completely separated if this is what you want. So we, we started with cd, but now we also have a CI product feature flag management, IDP that, uh, he mentioned mentioned chaos, engineering, cloud, cost management, and, and many other products. Yeah.

So all them are are separate product, which you can use independently if this is what you need, if this is your priority. However, if you use them all together under the same platform, they don't feel like separate product. They're all, That's really The key.

Exactly. They have the same user experience. They all utilize the same common components.

So it's the same pipeline, same, uh, access control, user management, policy engine, the dashboard. So when you use the harness platform, those products actually act as different features of the same product. So it's a great user experience.

Love it. I wanted to talk a little bit about, you know, probably, look, this is a great show and the cloud, we're gonna jump into cloud native software in a bit. But first I want to hit on, because we're five minutes and we gotta mention it, ai, right?

You can't go five minutes without talking about ai. Yeah. We're Already running late.

Yeah, Exactly. Well, in more ways than one when it comes to ai, but certainly AI is having a, an impact already on almost all the software we're working on. But I, I, you know, I think we also all realized that we've kind of just scratching the surface, maybe's a lot more meat on this bone.

But Harness recently announced some new agentic type ai uh, functionality. Let me, if you don't mind, and if you can talk about it, what, what exactly is it? Uh, Abs Absolutely.

So you, you're covering harness since Harness came out of Delta, you know, since we're actually using machine learning since the very beginning. This is not way before AI was so prominent. It was cool.

Yeah. Before it was cool, before it was easy, right? We embedded, uh, machine learning into our first product, continuous delivery.

Uh, we called it a continuous verification. And what we actually did was machine learning back in the day was correlating data from different sources to understand the outcome of your deployment. Because it was a continuous delivery solution, we wanted to help DevOps and platform engineers to understand what is the outcome of the release?

Was that a successful deployment? Did it fail and help them to fix it and decide what the next steps are. And then when we launched, uh, our second product, which was ci, we also used machine learning in a feature which we called test intelligence, which what this feature did was help you to reduce quite significantly the build cycle time by understanding which test should be executed whenever you have a new build.

So instead of running all your different tests, every time you run a build cycle, we understand which tests are more likely to find quality issues and fail. And we only run subset of your tests. So we can significantly, uh, reduce your test cycle up to like 90%.

And we saw amazing savings by harness. We're using this, uh, internally and across our customers. So ai, we're not strange to, to ai.

This is something that is part of our culture. But in the last few months, we uh, introduced, uh, three, uh, AI assistants. The first one is DevOps assistant, which help, uh, DevOps engineers and platform engineers to, uh, plan and build their pipelines, their DevOps pipeline in a much more efficient way.

The second assistant is code assistant, which help developers to write code in a more efficient way. So taking the way, uh, taking all the toil from the code writing and helping you to write code that is more efficient. And the third one is QA assistant, which help you to plan to create and execute tests in a much more efficient way.

Test has always been a very tricky part of, uh, of the DevOps ecosystem because tests are very brittle every time the code change, the test, uh, break if you don't maintain them properly. So we built an AI system that help you to understand what change in the code in the, in the code and fix and update the, the test, uh, properly so you don't have to worry about it. And your test will never break.

And it only updates, it does not delete the test to ensure it passes. Of course, of course. Yeah.

We ensure you have the same level of coverage. That's a good point. So let me ask a question.

Would you consider these assistants agents like Ag agentic ai, or are they more like co-pilot type of, uh, helpers? Yeah, I would say they're closer to the co-pilot. Uh, okay.

Correct. They're not really agents, we call them assistant. Okay.

They actually help you to be more efficient, whether you are are DevOps engineer, uh, you know, a QA engineer or a developer depending on, on who you are. And all three of these, they were announced, I think at the latest Tarn conference. Unscripted.

Yes. They're already available. You can try them out.

They're proud of our, uh, free, uh, free package so they are available to all our customers. Excellent. On it.

So that, that was the first You can get those. io, isn't it? Yes.

Harness io. io stuff? Oh, that's, so we're still trying to figure this out.

Yeah, it's gonna be interesting. I Been that you're only the fifth company today that I've asked this of. Yeah, because everyone's io who knew?

Yeah. Anyway. Alright, let's move on.

From the agent or from the, uh, AI assistance recently harness released its own open source. Yeah. It's kind of soup to nuts platform here, but a bunch of open source.

Correct. Tell, talk to me. Make us smart about that.

Yeah. Uh, you want me to cover that? I can.

I can. And you, you can, uh, do that later on so I can start. Um, sure, go ahead.

Uh, uh, you cannot do, uh, injuring like developer software without open source. CubeCon is the biggest open source conference and, and we are here, this ecosystem existed because everyone did Open source Backstage is open source, open Tofu is open source. So actually harness the core part of Harness Pipelines is built on one of the first open source CI platform drone.

Um, and that was our initial journey, I remember. Um, and, and, and then we, we kind of rebranded it, uh, to harness open source. So, um, harness Open Source is, is a collection of writing code.

So it has a good repository. It helps you build pipelines and then deploy, um, have an, uh, a small cloud development environment. Um, so it's, it's a great way to get started with harness, understand what is about to come.

Um, but then, um, when, when time comes to really make it enterprise grade, um, you know, you, you shift over to harness. So that's how I would put it. uri uh, no, Absolutely.

Yeah, you're right. You're spot on. So Harness has always been huge sponsors of open source, uh, project.

So we sponsor Backstage, we're part of the founding team of the Open, uh, TOFL project, which is part of the Op Linux Foundation. Yes. Uh, we are sponsoring Litmus, which is A-C-N-C-F project on gas engineering drone, as you mentioned, mentioned.

So we, we, we are big, uh, on, on open source. Yes. Uh, but last month we made a very exciting announcement.

We announced harness open source, right? Which is a complete DevOps platform. A hundred percent open source and very lightweight.

So this is great if you are a practitioner that's need, you know, lightweight platform that you can install on your laptop and get it up and running in like 10 minutes and it has everything that you need. It has cloud development environment, GI repository, ci, cd and artifact registry, all open source, all tied together Artifact, like Artifactory, Like Yeah. An alternative I would say yes.

An artifact. Very strict. So you can absolutely manage the entire lifecycle of your, uh, Kubernetes, uh, platform using this solution all from your laptop.

And this is what, I don't know how technical or deep you want to go here, but what, what os license, what OS is what, what, what Open source license. Apache, I think, right? Yeah.

It's Apache, Dr. Apache. It's Apache, yeah.

It's not a business license or anything. No real open Source. It's not a one off kind of No, no, no, no.

It's an OSI license. Real Open source. Absolutely.

Excellent. And that's put out, it's not A-C-N-C-F project. This is hard.

This is open source for people who just wanna get sort of a lightweight platform that gives you, you know, that DevOps platform. Exactly. It's, it's very lightweight.

How do people get that? Oh, the, the respond was, was amazing. No, No, but if someone out here says, says, I wanna go check it out.

Yeah. Harness said IO slash open source and that's it. Great.

Yeah. Alright, let's move on. Now both each of you are, you have a near and dear open source project.

Oh yes. That each of you work hard, you work it with, uh, it's Open Tofu. Correct.

The Open, open, uh, code Infrastructures Code. Correct. Correct.

And, And our audience knows Open Tofu is a fork of, of ha of Hashi. Correct. Uh, Terraform, I'm not gonna get into the politics in behind it, but it's been very successful.

The CNCF has done. Right. A good job in shepherding this, I think, my opinion.

I would agree. Talk to us about your involvement and why you're passionate about it. Yeah, yeah, absolutely.

So, uh, as soon as, uh, you know, the, the license for Terraform was changed, right? Uh, harness and a few other, uh, companies in, in the industry decided to basically sponsor a fork of Terraform, which is open Tofu because we wanted to make sure that the community have a real open source solution for infrastructure is good as code. And the first thing we did was actually submitting it to the UX Foundation because we wanted to ensure the community that this is going to be open source.

No one, no one owned it like That. Exactly. Exactly.

It's going to be open source for life. And, and the adoption has been extremely positive by small companies, by enterprise, by different vendors, by the providers. Um, we are seeing a great traction of, of open tofu because again, it actually delivered the promise of a real, a true open source, uh, solution.

Uh, also the migration from Terraform to Open Tofu is, is seamless. There's nothing you have to do. We can, we are fully at backward and forward compatible with telephone.

So at any point in time, we can switch from Terraform to open Tofu and everything will work as expected. Got it. Open Tofu, of course, is, as we said, it's one of the CNCF projects here, and It's a Yeah.

Open. We get Whole Nation correct. And Harness provides, uh, an enterprise solution on top of that, that help you to manage and orchestrate, uh, open tofu automations with governance and security and scale, which you probably need if you work in a large, uh, organization.

I love it. Excellent. Now, Mon, you already mentioned you had started with Backstage.

Yeah. Right. Which is also A-C-N-C-F project.

Talk to us about your passion on that. Yeah. Um, so developer portals are not a new thing.

Every company has a portal and I would argue multiple portals where they're expected expecting developers to go to do a certain thing. And people love building custom portals. People love, especially engineers love creating that custom experience for their companies.

However, it costs a lot. Um, you know, like typically we see, you know, if you've got a thousand, 2000 developers, you're spending five 10 engine capacity just to maintain that portal. Um, and then there's no standard.

So anytime you, you, the engineers leave, you have no idea, and the, the portal gets deprecated and now you have another portal to replace the previous one. This whole thing, um, has been a mess until 2020 when Spotify saw all of this and, uh, saw that back. Spotify engineers are using backstage every single day to be much more productive.

Um, so they decided to open source it. CNCF accepted it, the community accepted it. Um, when I was there, the, the project was getting about 200 pull requests and, and issues every single week.

And it was such a fast moving project. Um, so it has now about 30,000 thousand GitHub and, and, and scrape. Um, so that's how it started.

Where, you know, every company can realize that vision without having to do custom stuff. And, uh, like the worst is you don't have a portal and, and everything is just thrown at developers. Um, why do I am, why am I passionate about it?

Um, I'm a developer, I'm an end user developer. Anytime I would love to put on some headphones the right code, I still do that. Um, it, when I joined Spotify within the first week, I created my first project.

I built it, I deployed it, it was running in a dev environment. And that is not an easy task in, in multiple enterprises. You have to go through approvals and take a month or two to get your first, uh, you know, project done.

Um, that's, that's a very powerful thing. I, I felt good about it and I learned about how to work in this company, and that is exactly the Vision Harness has for our, our own customers that it should be easy in a large enterprise to, for a developer to come in and, and create a new project as fast as possible. So Backstage is there.

We have a product, uh, which uses backstage called Developer Portal, uh, which our customers can use to realize the same vision. That should be really easy to get a, get a developer onboarded, get a new service onboarded, and reduce their cognitive load, um, if they're working in a very large, uh, automation. So I, I like it here.

I think it's a great vision and, uh, yeah, we'll see more happy developers, um, in the world. I love it. Gentlemen, we are over time.

Oh, it was fun. Okay. We, we, but I think we covered what we wanted to cover.

Absolutely. Come on Uri. Thank you both for coming here on Text Drunk tv.

Enjoy the rest of Q Con. Thank you for inviting us. I hope you're loving the, our coverage today here.

You got the harness story. Check it out. We'll be back.

We're live at Q Con.