Last week, I had the opportunity to speak with J.C. Granger, host of The Future of BizTech Podcast (video here). We discussed the growing complexity of applications in the enterprise and how Glasnostic helps operators understand and control their behaviors. Below is the transcript, edited for clarity.
J.C.: Welcome everybody to another episode of The Future of BizTech. I’m your host, J.C. Granger. I have with me here Tobias Kunze, who is the co-founder and CEO of Glasnostic. Tobias, thank you so much for coming on the show. Tell the audience a little bit about yourself and what is it that your company does?
Tobias: Absolutely. Thanks for having me on the show.
I’m a Platform-as-a-Service guy. Previously, I was tech co-founder of the PaaS company that became Red Hat OpenShift. So I spent a lot of time thinking from a technical perspective and from a product perspective: “How do you optimally support the building of applications?” But seeing OpenShift run as a public service, it became apparent that building applications is no longer what’s difficult. It is the “How do I manage the code once it is running?” part. That’s where the rubber hits the road—the operational side, and more specifically, what’s called “Day-2 Operations.” At Glasnostic, we focus on what happens after you deploy the code and before you deal with incidents. In that window of time, there’s very little we can do. Once something goes wrong, the only thing that we can do is roll back or maybe reboot machines and that kind of stuff.
And it’s actually really surprising how much manual labor there is. There’s a lot of monitoring going on and a lot of looking at data—deep data—and root cause analysis that takes hours or days. Then there’s the incident management piece, where you need to get all the teams together. You need to decide what you are going to do. And meanwhile, your customer experience is nose-diving.
We automate a big part of Day-2 operations. We say, “Don’t go deep on the monitoring side. Just look at how the systems behave and manage what’s going on.” And then you can still decide how much of a manual process you want to attach on the backside of this. Meanwhile, the situation is in control.
So our value prop is really “Don’t sweat the Day 2 Ops piece. We’ve got you covered.” We automate it and give Operations, DevOps, and the DevSecOps teams real-time, runtime control over what’s happening at a very high level.
J.C.: That’s great. So then tell me about what types of clients do you have? Are these enterprise Fortune 500s, are these small business owners who are developing apps or SaaS of their own? Which type of client can you really make an impact for?
Tobias: Great question. We thrive on complexity. Glasnostic is in the business of massively reducing operational complexity.
If you have one application, that’s not complex. It’s like building a single-family house: you can build it however you want, that problem is solved. Where we come in is when you have something large, industrial. When you have a shop floor, different production lines—in short: complexity. When you have to accommodate a lot of processes that depend on each other. Then you need a supervisory function that is essentially a management layer on those operations.
Put differently, if you build something small, you’re flying a single-engine plane. If the weather’s nice, you don’t need air traffic control. If there’s a smaller airport somewhere in Nevada and there are 10 planes a day, you just look out the window, announce you’re about to land, and you land the plane. However, if there are hundreds of planes in an airspace, you need air traffic control because everything becomes unpredictable. You need to make sure everything is controlled and the airspace remains stable.
That’s what Glasnostic does.
J.C.: I like that analogy. You’re air traffic control for complicated business setups and systems. What’s an example of a typical client you have?
Tobias: We work with Fortune 500, Fortune 1000, Global 2000 enterprises with complex software landscapes, generations of systems, and tons of integrations. That’s where we shine. Conversely, if your applications are mostly siloed—even if there is scale, but your applications are not connected—that’s not that interesting for us. These apps are not hard to write and, most likely, will be just fine.
J.C.: What industries would be more complicated than others where Glasnostic can help?
Tobias: We are pretty horizontal as a product, so we are not glued to one vertical or another. If I had to call out a couple of segments, I would say managed service providers are a significant segment for us. Simply because, as an MSP, you’re running other people’s workloads. That code is not yours, and you need to protect yourself against whatever these applications are doing. Also, how you run these is actually already pretty complicated. I’d also call out financial services and automotive, and retail. These are all sizable operations, and there needs to be a sizable amount of systems for us to bring value.
J.C.: What is your company doing marketing-wise to get the word out about what you do? How do you guys reach out to other companies, so they know you exist to provide this service?
Tobias: We are still early but are already pretty busy with referrals from customers. I like to blog, so we’ve published quite a few thought-leadership posts about how to operate something that is fundamentally too large for any single person to comprehend. Those get good inbound interest. Also, I enjoy being on podcasts, and we have a lot of word-of-mouth referrals at the moment.
J.C.: That’s a great spot to be in. How do you prioritize new features and releases in your software?
Tobias: That’s a difficult question because there’s just so much for us to do! We’re growing, and we’re constantly hiring, but of course, what we do is also deep technical, so it takes time to onboard people. In short, we prioritize pretty aggressively. We are still early as a company, and fortunately, many of our customers are also early in their journey to automate ops, so there’s quite a bit of understanding and compassion from our customers, and they are happy to evolve with us.
J.C.: Personal question: what did you want to be when you grew up, and how did you get to where you are today, with Glasnostic?
Tobias: It’s really funny because I wanted to be a musician!
J.C.: For people who are only listening and can’t see, he’s got this marvelous grand piano in the background.
Tobias: That’s what I still do today when I have time. But yeah, I studied composition and conducting. From there, I went into computer programming. At first, I started out doing digital sound synthesis, which involved quite a lot of programming. And yes, I believe composing and programming are really kind of the same thing in many ways. Both are fundamentally creative activities. You need to think creatively to come up with good solutions. So from there, I worked in a couple of large companies, and then have been in startups ever since.
J.C.: Where do you see your industry going in the next 5 to 10 years?
Tobias: My view of where things are heading is pretty unique, I think. For instance, I believe development is getting easier by the day. These days, you’re mainly developing narrow, bounded-context services that are a couple of thousand lines of code, perhaps. You’re not building applications with multi-million lines of code, but you’re building many small services. We’re in a world today where we can actually develop very quickly, and the developer experience is getting better by the day. We can build with many excellent technologies and tools, but now we’re running way more pieces in way more places. It’s not just on premises and cloud. It’s also hybrid, multi-cloud and edge. And in addition, we also increasingly connect these systems, so part of the result of writing small services is that I’m going to combine these with existing capabilities that I already have.
J.C.: It becomes a Frankenstein model in a way,
Tobias: Yes! Just think about it: in any organization, how do you assemble teams? You pull a couple of good people from other teams and say, “Hey, we have a task force here. Can you do this?” That’s how new applications come to life. You compose and recompose. And that’s kind of the model that’s taking hold in the industry. It started a little bit with microservices. But most engineers still think of microservices as a blueprinted, fully designed application, kind of like your single-family house. You have the floor plan and a couple of variations, but the house doesn’t really change. But because of the pressure from the business side to accelerate, companies are splitting up development across parallel teams that all deploy autonomously. They can all push code into production on their own schedules. And now that code joins all the other code that’s already there. So you get into a massive, multi-body physics problem.
This has two important aspects. Number one is: how does your transaction execute through all these systems? Are there any bugs, anything that’s taking too long? Is there anything on the path of execution that’s not right or should be different? That’s one big chunk of problems. But I believe this set of problems is largely solved today with the existing tooling we have. There’s a lot of observability and monitoring out there.
On the other side, we increasingly have these strange environmental behaviors. Something unrelated gets deployed, and all of a sudden, my dependency behaves a little bit differently, which can throw me off. Or, all of a sudden, something happens, and I don’t even know that it happened or where it came from. Some limit has been reached in some component. Maybe a connection pool that runs out of space, or a message queue that’s full, or something that takes too long, so a related automation is triggered, and so forth.
We live in a highly automation-rich world today, and none of these automations are designed to work together. There’s a lot of impedance mismatch. So that’s where we come in. We are not looking at the execution of code at all. That’s the pilot’s problem if you will: how do I land this plane? How do I get it to the destination? We are looking at what is happening in the airspace. No matter how well you plan and how much you automate—ultimately, unpredictability reigns supreme. There’s a lot of “unknown unknowns,” and the key is to detect these quickly enough and react to them immediately.
Importantly, though, reacting does not mean trying to hunt down a root cause. That often takes hours and days, and sometimes there isn’t even a root cause at all—it’s just a confluence of unrelated factors. What’s imperative, though, is to bring the situation under control and create predictability—to stabilize the patient and stop the bleeding first. Only then do you start a full-blown incident management process, and only if you believe it will be worth the time.
So, that’s kind of the story of what we do, but that’s also my prediction for the future of our industry. We’re going to run more and more things in more places, and it’s very clear that there is a natural limit to how much we can achieve with better thread-of-execution engineering. Therefore, we need to create resilience in a different way.
J.C.: What kind of new features or services could the audience look forward to?
Tobias: There are quite a few directions we’re working on. One direction we are spending quite a bit of time on is to detect more interaction patterns, more anomalies. Not every pattern becomes actionable, but we want to know about the most significant threats and flag those patterns. The other direction we are investing in heavily is to increase automation. Automation doesn’t necessarily mean a fully closed-loop design where you just let the machine do whatever it thinks should be done. It can be delayed-action automation, where you can observe and think about if you want to take action before a default remediation is applied, and so forth. Often, the human brain is much smarter than any machine—especially in operations, because context matters a lot—but we can use automation to make it much easier to manage these large exploding infrastructures.
J.C.: My last question for you is, what’s the best piece of advice either you’ve been given or that you can give from your own experience to the audience?
Tobias: I feel it’s been given to me but don’t remember who gave me this advice: when I studied composition, I learned that we tend to underestimate how much creating something actually changes the world. So I feel a lot of my peers are focused on the competition and on what the market is doing. That thinking puts you in a reactive mode instead of focusing on what you can do. By releasing a product that solves maybe even just 50% of a customer’s need, you put a stake in the ground and create a reality that you can use to take a second step. This is what matters, not the competition. That’s something that I wish we would focus on more in the industry.
J.C.: To further your point about people reacting to competition, I used to say to people, “I don’t have a rearview mirror.” I have no idea who my competition is because I don’t care. If I’m reacting to them, that means I’m a step behind them. I prefer that they react to me. We’re always trying to improve in our own processes. As long as we’re doing that, we’re looking forward, and we’re not looking backward. So I like that advice.
Thank you for coming on the show. How can people reach you and Glasnostic?
Tobias: Very simple. The company webpage is glasnostic.com, and my email is “tobias” at “glasnostic.com.”
J.C.: Thank you again for coming on the show.
Tobias: Thanks. It was a pleasure!