Gremlin CEO Kolton Andrus on Chaos Engineering, Observability, and AI in Product Leadership

Hey, if you've watched Tech Drunk TV over the last few years, you may may have seen this gentleman on his name is Colton Andrews. Andrews Andrus. A-N-D-R-U-S.

That's correct. Not Andrews my funny French accent. You gotta be careful with that stuff, Colton.

But Colton is the former once and present and future CEO of Gremlin. That's right. Right.

And um, if you don't know who Gremlin is, stay tuned, we'll tell you. But Colton, it's great to have you in person. You usually do this, you know, remotely over text, drunk tv, but here you are in the flesh, man.

Yeah. Good to see You. Remote's fun, but, uh, different energy when you're live So much.

Enjoy. You know what I, um, I gotta echo that, right? I, I just had this at a, we were at a dinner last night and I was talking to some people, whether it's a, a sales environment where you're there talking with a potential customer or a partner or even a friend.

Nothing, nothing takes the place of person to person. And, and that's not, and that's something AI's never gonna be able to do. Yeah.

Right. Such a timely comment. Yeah, exactly.

No, but humans are social animals. And though, like when we were locked in our houses for COD and everything else, zoom was the next best thing we had. It pales in comparison Yeah.

Of, of this kind of communication. So, hey man, it's great to have you here. Yeah, Yeah.

Thanks for having me. So, Colton, I made a little reference to once present and future CEO, but there's been some gaps there. Talk, tell, give people a little bit of your history.

Yeah. Well, uh, for those who don't know me, I'm Colton Andres, engineer by trade. Uh, grew up working for a bunch of tech companies, had an opportunity to work at Amazon for a few years, focused on their reliability program, got to join Netflix, got to continue my good work there.

And that really led to going out and founding Gremlin. Uh, first few years I was CEO, uh, for six years, you know, out building the product, learning the market, really advocating champion for the idea. Uh, brought in another CEO to run the go to market side so I could go focus on product and engineering.

And that was really a go into the lab, you know, understand the problems well and go fix 'em. And, uh, I got that done. Engineering's running great product's in a good spot.

So I took back over as CEO and excited to be, you know, running the whole ship again. You know, look, I know a lot of guys who were, you know, founded companies, became CEOs and then moved into product role, CTO role, chief evangelist or strategy. But I, I'll be, I'll be honest with you, I haven't seen many come back to be the CEO.

Yeah. Um, so kudos to you on that. You know, if you wouldn't mind, you know, I don't want to go Barbara Walters on you showing my age even saying that, but what, what kind of, what was, what was like, what was the light that went off and said, okay, it's time for me to come back as CEO?

Yeah. Well, I think it was twofold. I mean, just to be, just to be completely honest, you know, if it's your first time being CEO, you have some doubts.

You wonder if you're doing things right. If you could have done things better, Joe. Right.

And so, you know, you bring in another CEO, it's an opportunity to learn, listen, observe. Uh, and I'd love to say, you know, that the person I brought in taught me a bunch of things I didn't know, but a lot of it was actually, you know, kind of par from the course. You're, you know, I was doing it.

Okay. And You needed that reinforcement. Yeah.

And so that was part of it. Uh, and then the, the product and engineering side, I'm an engineer, I'm a product guy, I'm a builder. And one of the mistakes I made was stepping too far away from the engineering side and really delegating as, see what you mean as CEO.

And so, yeah, my, my three years as CTO was getting back in and making sure we were building product correctly, we were understanding it. Well, part of it was getting our engineering team just humming, you know, making sure they were shipping. It's the lifeblood of a tech company.

And if you've got issues that are preventing, you know, consistent quality delivery, those have to be fixed. So once I had those fixed, that left me in a position where I felt like I could come back and take back over the reins as CEO. And I think one of the, one of the big things for me is one of my roles has always been an evangelist role.

And it's actually a bit easier in the CEO position to be that evangelist. Sure. To be in front of customers, to be in front of prospects, to be at events, to be doing interviews with great guys like you that allow you to get out and spread the message.

And so that was, that was a big impetus. Flat, flat flattery will get you everywhere, my friend. Thank you.

But you know, I, I had a little bit of the opposite from you where, and, and this is another sort of, uh, persona that I, I've seen in startups where CEO founder is the chief sales guy. He's the guy who closes deals. He's the guy who has the network, or she, in some cases, right, they're the ones who are primarily responsible for making it rain.

And sometimes, you know, revenue is key, revenue is king. And you're so focused on trying to go out there and make it rain, bring in the revenue that you, you, you lose touch or you lose focus on is is product running, right? Is engineering, is operation now a case, it's operations, it's, it's, you know, all the little things that beyond doing these, 'cause I'm the face of it, beyond, you know, making it rain.

And, and that's a, it's a tight, you, so you live this too. You're just living it from the engineering side. I'm living it from the revenue side.

Right. You, it's a tight rope between how much focus do I internally give to the, the guts of the company versus my external focus on making it right. Yeah.

Yeah. And that's, that's one of the parts. Yeah.

I think one of the unique positions I'm in is being an engineer that has done it and lived it. I have a credibility, I have a, a wealth of experience I can speak to and that's relevant whether I'm talking to engineers or I'm talking to executives. Yeah.

And so, but yeah, if you get, you know, gotta gotta have revenue, gotta build a healthy company, gotta make sure the company's moving forward and that, that focus can steal from kind of the pure engineering. Are we solving the problem well and are we really making the progress we need to? Absolutely.

Tight balance. Absolutely. It is, it is a tight balance.

And, and the other thing is, and, and I would say this is probably another mistake that CEOs make, especially first time CEOs or more inexperienced CEOs, is you hold on to things so tight, right? I, I can't let go of engineering. I'm, I'm an engineer.

I, I know I have the vision. I God will see that too. I'm the best advocate evangelist we have.

I I'm the one who could close these deals. I'm this, I'm that I, I, I, I, you can't be all eyes. There's gotta be some whe in there.

And so learning to delegate, first of all, it's What's, It's having a team, you could trust It. It's interesting you say that because I actually think one of the traps I fell into was the too much good advice trap. Oh, You delegated too Much.

And I delegated too much. I tried to find the perfect executives. That's the antipater.

And a lot of what I've been learning and doing the past year is just falling back to what I find. You know, there's a balance. And so yes, I think what you're saying is totally valid.

You gotta build the team, you gotta build the expertise. But sometimes if you over delegate or you do too much, then the company goes astray. It's, it's, it's wandering too much.

There's too many options away from You To go. Right. And so I think that's the beauty of being in both roles and being in both positions is I arrived at this happy medium, which is, I have the confidence that I know the direction we need to go.

I have the confidence in the pieces I can do to it, but I have a team I can trust that I can delegate to, to ensure that I don't have to be the one who does everything. Love it. It's, you know, it's an interesting thing.

We're here to talk about Q Con though, but this is great. Look. Sure.

com. They'll be happy to answer 'em for You a hundred percent. Um, But Colton, let, let's talk a little bit about Gremlin.

You guys kind of, I mean, for in, in many ways you invented, well Netflix did, but you invented chaos engineering and of course you were there at Netflix. Mm-hmm. When chaos engineering kind of took root and gremlin, at least for my money, was the Chaos engineering company then.

Yeah. Right. Now the world changes, things changed.

Technology changes, but talk about today's gremlin. Yeah. So I think the Chaos Engineering Gremlin was a cool idea and a cool project, but really needed to grow up a little bit to be effective in today's enterprises.

I think a lot of what was great about the early product is it lets you go out and experiment and test and understand. And it was a bit how an engineer would build a freeform, figure it out. If you need help, come ask us.

Well, most enterprises don't have time to go out and figure it out, and they're busy. And so what they need is a bit more guidance, a little bit more prescriptive approach. So a lot of what happened over the last few years is let's take what we know the best practices are and just build them into the tool.

So people are starting out with the right patterns. And some of that is, uh, you know, the engineers aren't sure what to do. Let's tell 'em what to do.

Let's tell 'em what test to run. Let's integrate with their monitoring. Let's tell 'em if they passed.

Let's, let's take the homework away so they can get the work done and get to the answers they need. The other thing that Chaos Engineering struggled with is how does the company view it? And is the company view it as important as security?

Is it something the company says, yes, this is something everyone should do? Or is it an unproven idea? That sounds cool.

And I think we saw a lot of sounds cool, but we've seen over the last few years in the way we've taken the product, a lot more companies buy into the, this is how it should be done for our company. And one of the biggest parts of that, one of the things I love to ask, uh, my customers, I've asked it three times today, is if you prevent a large outage, can you get promoted for it? 'cause we know you can get fired if you cause a big enough oneand.

And the answer is, if you don't have some way to measure it, some way to track it, some way to prove it, then the business isn't gonna believe you. And that's really a lot of what we've filled the gaps in. Do you, do You think it's a question of the business doesn't believe you, or the business doesn't value it?

Kind of the same thing in the way I'm talking about it, which is, you know, if it's, if it's a clear value to the business, saves engineering time, it prevents customer pain. It prevents us from losing revenue. If they have confidence that our solution gets them, that we are, we're aligned, we're off to the races, we can get a lot done.

It's when people aren't sure. And we really have to go out and prove it that, you know, we got work to do. I agree.

I agree with you, man. Um, interesting stuff. io?

com. com if You want. Nice and easy Information on Gremlin.

Nice. Well, everybody tastes got io ai T that's my series A money went to our domain. I, I It's good.

Best One More. First time. CEO advice here from Colton.

Um, let's talk CubeCon. Yeah. You guys, you guys made a big announcement.

Yeah. Talk to us. Yeah.

Excited to announce our Dynatrace partnership and integration. We got a lot of great monitoring tools, but Dynatrace is one that really gave us what we needed to make it easy to do the right thing. This is something you'll hear me say a lot, make it easy to do the right thing.

And when it comes to this kind of testing, you really have to know how's the system behaved? Did it work well? Or else you're flying by blind, right?

You know, and so Dynatrace makes it really easy for us to find the code, the code that we're testing, we can go find the monitors, automatically pull 'em in and associate 'em without the engineer having to go do a bunch of legwork. Uh, and so it allows us to, uh, one, create these services within Gremlin based on what's already created in Dynatrace. So we can just pull it into Gremlin, set up your services, set up the monitors, and now day one, you're ready to just click run and go get answers instead of doing this homework you have to do in advance.

Absolutely. You know, we, we work a lot with Dynatrace. I mean, we've had Dynatrace on here a number of times.

They really got their act together in this observability and space and everything. They, you know, Dynatrace is an interesting company as well, right? They, they're not new.

They've been around, they were originally a European company. They bought a big American company. They then merged and there was a little, I think for a couple of years back then a little struggle over what culture and what vision they were following.

You know, we didn't have something called observability then, right? Yeah. It was a, a, uh, A-P-M-A-P-M.

And um, but now they've emerged over these last couple years with their act together, certainly. And they're killing it. I've seen a lot of great product innovation.

Yeah. And there's things that, you know, beyond just alerts and monitors, you know, when it comes to reliability, you need very fine grain data. This is something where we need to know what's happening every second, not every five minutes.

Right. And we found that granularity. We found the nice places we could integrate to get that visibility.

Yeah, no, I've, I've been impressed with the product innovation and, and the partnership. Absolutely. Let me, let me put on my CEO rainmaker hat again here.

All Right. I love it. How, how is this translating in the market for you?

Or how do you expect it to translate? 'cause it's new? How do you expect it to translate in the market?

Well, I think we see a lot of, uh, shift in the observability market, uh, from our position. A lot of companies are reevaluating their vendor of choice. Uh, cost is a big issue.

Sure. And so people are trying to right size and find that value. And the truth is, we are, we're pretty tied to that observability.

If you can't see what's happening, you really can't do intelligent testing. Agreed. Uh, and so this ability to really understand what's going on, integrate in and get that helps us to be able to just move so much faster, be able to get that value so much quicker.

Agreed. Agreed. So, yeah.

Is there, is there plans of how you go into market here though? Is it, are they gonna sell this solution or offer it or talk about it, or, I mean, right now it's an, it's an integration and it's something that shows up in our product. We've got some things that we push into their product as well, so people can know what tests are happening and when they're occurring, they can correlate 'em with events.

Um, but for us, we just, we have a lot of important Dynatrace customers. We wanna see them be as successful as possible. And so leaning in with Dynatrace lets us go build a better product for our customers, which leads to our customers getting better adoption, better reliability results, customers.

That's the flywheel we want to get going. Love it. Other observations around coup car?

I realize it's only day one. Well, if you were here yesterday, right? There was a lot going on, but what, what, you know, what do you think?

It's been interesting. I've, I have only been able to walk the floor a little bit. Um, you know, I think the last conference or two, you were kind of getting beat over the head with ai.

Yeah. I feel like it's balanced out just a little bit. Not every single person's talking about it.

Uh, you know, I thought it was a, a compliment. I got, uh, a customer came by the booth, uh, or a company, a customer company, not somebody I'd worked with directly. And they asked for the details.

We told 'em all about Gremlin. They're like, great, we want this. And then he is like, you win because you didn't swear you didn't say the, the curse word at me.

And I said, what, what curse word? He said, ai. And I said, well, you know, there's, we've got some intelligence stuff.

We think it's important, but that's not, we're not, we're not here to just jump on marketing bandwagon ai. It's The answer for everything. It not, not to say that it's not real, it's not disruptive, it's not all of those things, but it's, it, frankly, it's not the answer for everything.

Yeah. I think there's great uses and places for it. I'm an engineer, so I'm a little cautiously optimistic.

I want to see the value be pr it's the same as chaos engineering. I wanna see the value be proven. Yep.

Before I go all in. And look, we're, we're building intelligent capabilities that make it easy to do the right thing. Well, it's, you run the test, you know how your system responded.

Well, the last step is you gotta go fix it. So we built a great product this year that tells people how to fix the things they find. And to us, the key is it's credible.

It's it's accurate, you know, and it provides good advice. And, you know, we're not, we didn't just slap it on everywhere. We put it in a very specific spot, we tuned it well and we feel really good about those results.

But, you know, the dirty secret is we're not just shelling it out to an LLM, you know, we're, we're doing some machine learning and modeling in the background and treating it a bit more like a, a classic data science AI problem than just a, Hey, could the LLM solve this problem for me? I love it. Hey Colton, we're about outta time.

com. That's the important thing to remember. Good luck for the rest of CubeCon this week.

Um, I guess we'll see you on Text Drunk TV next. Yeah. Yeah.

I'm looking forward to it. Thank you very much. A Colton Andrews CEO of Gremlin here on Text Drunk tv.

We're gonna take a really short break 'cause I got my next victim sitting in the barber chair. Ready to go. You're watching Text Drug tv.

We're live at CubeCon.