Revolutionizing Cloud Monitoring with Cloud Canaries’ Mark Callahan

This is Textron tv. Hey everyone. Welcome back here to techron tv.

I'm really happy to have a first time guest here on Techron tv. He is the co-founder and CEO of a new company called Cloud Canaries. Welcome, Mark Callahan.

Hey, mark. Welcome. How are you?

Thank You very much. I'm, I'm, uh, uh, it's a pleasure to be here. Absolutely.

So, mark, are you're based up in the Cambridge, Boston area? We are right between MIT and Harvard and Sure. Cambridge.

I, as we were talking off camera, I actually helped start a company that eventually we iPod that we were in Kendall Square in the building. I think it was Ben Franklin or someone was on the roof, like the printers building or something, it's called, um, fun times. Good stuff.

Mark, before we get into Cloud Canaries, if you wouldn't mind give us sort of a brief kind of history or story of your journey to becoming the co-founder and CEO here. Well, it, you know, it, it all started out MIT as a lot of these stories go, um, I was fortunate enough to be part of the kind of the early years of, of what turned into the media lab. Sure.

Uh, it was some of the best time of my life. And from there I kind of went from startup to startup. And the first one was doing, uh, um, um, engineering systems, um, software, uh, for a company called draw based software.

And, uh, we were competing against a, a company that I wanna kind of em emulate. And that was, uh, Autodesk and their product AutoCAD. Okay.

And from there, some additional startups. And the last one before Oracle purchase us was Profit Logic. And we did demand forecasting for retail telling retailers what was the optimal point in time to actually sell out all your merchandise.

And we, uh, were lucky enough to, uh, virtually have all the, uh, uh, high fashion retailers use our software. And, uh, eventually all the, um, uh, Oracle purchased us. And, uh, from there I was with the kinda retail engineering and, and support organizations for a handful of years.

And, and at some point, I, I got lucky enough to actually be part of their, uh, cloud, cloud work that, um, they spun up very quickly and, and it was, it was very, very interesting. And that's actually where I actually got involved with canaries, or at least learning about them. And canaries are the canary in the coal mine, but our canaries are, cloud canaries are very smart, don't die.

Um, and, and what they do is provide monitoring information. Um, they observe, they can be used for testing, they can be used for virtually anything. Basically they create work workloads.

And with a workload you can get, uh, data back, uh, status information that we can then take and feed into a neural network and actually get forecasting and predictability out of it. ai is actually part of that. Got it.

So, you know, the, the, the concept of a canary in the coal mine is not a new concept in tech. Right. And specifically in DevOps, for instance.

Um, it's a way of ab testing, right. I, I forgot that it's an open source project, I believe, uh, red Canary or something like that where look, you, you put out a couple of different versions of, of a given Apple whatever with some minor variations and, you know, feature flags and stuff like this. And, and you just, you see which, which performs better.

Right. And then you obviously, you know, it's kind of Darwinism in real time. Um, it sounds like you've taken that concept and extended it beyond just testing, you know, ab testing features kind of stuff, uh, or, you know, feature flags into really monitoring, uh, monitoring just about anything.

'cause you could put these canaries or agents if you wanna call that Yes. Anywhere. Yes, it's exactly right.

Our, to us, um, a cloud canary is a very simple piece of Python, Java code even go code that basically creates a workload and then gets that information back from that workload status information and other pieces of data. So with that information, we can, um, a determine if that workload, that service is healthy, uh, but also generate forecasts to make predictions about that service are groups of services or groups of APIs all from a set of canaries that a developer can throw together, maybe not throw together, but develop in 45 minutes, deploy today and predict tomorrow. Got it.

Um, so Mark, is current canaries to be used primarily pre-deployment or is it post-deployment? And then, you know, you take that feedback for iterations? So it can be, cloud canaries can be used in both worlds.

It certainly can be used in as you begin to roll out, um, software in, into the cloud. Um, so that's, that's definitely a given. But you can also, and I think more importantly, use it for continuous monitoring observability and also leverage it for a, a spectrum of, um, additional solutions, whether it's life sciences, whether it's security, um, you know, digital experience.

Um, so I think the bigger win is actually in production. Okay. Um, so it's to the right of the deployment, uh, you're, You were be you're gonna have cloud canaries both, um, filling, um, targeted canaries running during test and, and integration.

Okay. And, and keep those canaries running, um, for a lot for, you know, indefinitely to basically monitor those, um, those APIs, those services. But the key here is also, it's not necessarily the developers who can use cloud canaries, but it's actually the users of those applications who basically can verify that their applica, that the application that they're paying money for is actually working to Be working application, Whether it's SLA, whether it's health, you name it.

Yep. So, mark, my next question is does in term, you know, an organization may have many, many canaries Yes. Out in the field, so to speak.

Um, is there a central Canary monitoring coop or something like that? Or like, does each canary kind of have its own specific UX or what have you? So a lot in the past, uh, canaries would just be a run Byron job.

Uh, cloud Canaries actually has the Avery platform that allows you, allows the DevOps engineer to create the canary to schedule it, to deploy it, to collect the data, to analyze it, to see the purdy graphs, but also to feed the, uh, neural networks so that you can generate a model and then do predictions. So we, you're getting both actual data in from the Canaries, but we're also generating forecast data, and then we can set our thresholds for alarming and notifications external systems to tell them that there might be a problem or something good that's happening. So it is a platform that, that really provides the canaries, the cloud, canaries, the intelligence to do what they do.

Okay. Um, and then I, I realize company just launched. Yes.

Yes. We, we launched, uh, on the 16th of July. Okay.

Congratulations, you on that. Thank you. Thank you.

It was a lot of hard work to get. Is is the product in GA at this point or? No, we're in production, so, oh, you are So Full on.

ai, go to, you know, you can go through the pages, uh, go to our pricing, uh, page and actually pick a subscription. We have a three month subscription right now that's free and, uh, that gives you, uh, um, uh, three canary licenses, or you go to our corporate, uh, subscription, which is 45 licenses, 45 licenses to run 45 canaries together at the, at one period of time. You can spread those out within our organization.

But the key here is, is that because we use AI and compute and workload data, we can charge literally a hundred dollars a month for those 45, uh, licenses. And this is where we've kind of shifted the paradigm where it's, it's really observability without instrumentation. Yeah.

Yeah. That is, that's a good way of looking at it too. 145 agents for a hundred bucks.

How, how big do you think the average deployment would be? How many canaries? So, so the way our, um, um, canaries are laid out within a company is by organizations.

So an org, uh, a large enterprise can create multiple organizations, and we figure about 15 canaries are running in each organization. And, and it could, the organizations are whatever you wanna define them. They could be development, test, uh, stage production, they could be marketing, they could be sales, they could be engineering.

However you wanna de define that. And, and each of those, I would generally think about 15 canaries can be running at any point in time that we'll actually be generating data and that we can, you know, do our, uh, modeling and forecasting, uh, out of that. Great.

ai, that's the website. Yes. ai.

How can people sort of get engaged right now? Someone says this sounds good, I want to give it a try. So it, first of all, we supply sample canaries that can be used immediately.

We can also provide, uh, we also provide, um, classes and workshops to allow DevOps teams and individuals and developers to actually create their own canaries. We believe that 80% of all canaries that are created are very similar between one enterprise and maybe market from another. So we want to allow, um, developers of these canaries to be able to share between DevOps teams for free.

And so we actually have a marketplace that allows you to take, uh, and download a shared canary and, and run it. And it's within our repository. We do the scans, we do the testing also, uh, we have a subscription, um, that allows you to pay for canaries that have a little bit higher, uh, value.

And, uh, it's a royalty based, uh, process. So you can then, you know, for another 10 or $15 use these canaries. That's very cool.

So the whole thing sort of sounds like that, that sounds like, why hasn't someone done this before? You think, mark, That's a good question. And this is where I go back to one of my favorite, uh, companies that I, that I actually competed against, um, um, Autodesk.

And they started out, um, competing against a hundred thousand dollars workstations, engineering workstations. And, and, and they and IBM came out with the IBM at, and they came out with a, a software package, I think it was under a thousand dollars, uh, that did basically 80% of what that $120,000 workstation would do. And what it did was create this huge ecosystem of engineers, developers who actually built tools for the platform.

So our platform is highly customizable, it's configurable, and it connects. We like to think to virtually anything. So developers can develop their, their canaries on our platform, their solutions, and, and the, the, you know, customers will only have one platform to manage.

And I think that's one of the problems we have now in DevOps. You know, there's, you know, there's a platform for anything, but DevOps teams are just always under a lot of stress and to actually manage all these different platforms that are supposed to be helping them. Dude, I get it.

Hey, mark, we're about out time, but you know what, first of all, congratulations on Thank you on starting the company. It's no easy feat for anyone out there who's never founded and watched a company. You know, it's a lot of fun, but it's a lot of work.

So congratulations and success to you with this. Um, it's gonna be interesting to see how this plays through. Let's, you know, I'm, I'm interested to watch it grow and watch where it goes.

Come back and keep us posted. I will, I'll be excited to do that. Thank you.

ai. Right. And it's C-A-N-A-R-I-E-S.

That's, that's the real spelling. Good. All right.

Hey Mark, good luck and thanks for coming on. Text Trump tv. Thank too, Alan.

All righty. We're gonna take a break here on Tech Trump tv. We'll be back in just a moment.