Eliminating Blind Spots in OpenTelemetry with groundcover’s Shahar Azulay

This is Textron tv. Hi everyone. Welcome back here to Techron tv.

Our next guest has not been on Textron TV in a very long time, more than a year, maybe even two. I want to say hello to Shahar. Azule.

Shahar is the CEO of a company called Ground Cover. Hey, Shahar, welcome back. It's been a while.

I hope all is well with you. Thank you. Great for, uh, great to be again.

Thanks for having me. My pleasure. So, Shahar, before we get into ground cover and all the things we want to talk about, I wanted to spend a little bit of time talking about you.

As I said, you haven't been on in a while. Give people a little bit of your background and how you came to be here. Yeah.

So, uh, I'm, uh, shahar. I'm the CEO of, uh, ground cover and, uh, you know, uh, an engineer most of my life basically. So basically, um, you know, as, as, as a background, I've been in engineering managering positions of, uh, you know, one sort of another for most of my careers for, uh, um, cybersecurity for over decade, and then almost a decade in AI ml.

So, uh, in the context of observability, I was, you know, using these systems, um, you know, systems like ground cover from all sorts of sizes, you know, choosing these technologies for my teams, uh, working with them on a daily basis. So, uh, you know, I know this, uh, this domain is a user. Um, and, and maybe that's the reason I was pulled in so, so hard to try and know, make a dent in this universe.

But, um, that's kind of my background. So I think, you know, as a CEO right now, um, uh, I, I think I know the pains of our users and where they're coming from, um, and, you know, it's a mission critical domain of their ability, right? So, uh, making sure that we, uh, we know how to address this is, uh, maybe an advantage for, you know, me and the rest of the ground company team that kind of knows these solutions for a while.

Sure. You know, that's such an important, um, especially early on in startups, right? That the CEO, that the executives understand the, the problems that the solution tries to solve under, you know, that they walked a mile in the shoes of the user.

'cause if you ever walked a a mile in those shoes, it's very hard for you to craft a solution that is, is on point for what that user's experiences and what it is that, that problem you're trying to solve. And then I think one of the issues you get, and I, you know, I I've been a startup guy for a long time myself, one of the issues you get is as the company grows, as the executive, you get further and further away from that user, right? And what his or her, that's true.

Were there everyday challenges are, and jobs are, and experiences are, and you, you kind of abstract it, you think, you know, based upon what you knew five years ago or three years ago. But it's changed. And, and that's why I think, and you know, it's, it's so important for CEOs, CTOs, CPOs to go talk to real people, not not your engineering team, not the Gartner analyst.

Not that you shouldn't talk to them, but it's important to go talk to real people, real people who are, have the problems that you're trying to solve and understand, you know, what their anguish is, what the problem is, and how close you are to solving it. It's an important thing. Yeah.

I mean, For anyone. Yeah, I totally agree. I think that specifically in the absorbability market, you know, um, uh, it's, it's a very, um, busy market in, uh, how it feels to the user.

But basically, you know, there's, um, you know, a short handful of, of, uh, solutions that eventually kind of penetrate your, the consciousness of, of people. Um, and the solution of, you know, how you can bring better value is very integrate, right? It's not exactly, uh, you know, just innovating on UI or something like that's a very deep domain.

Um, and you have to understand the advantage and disadvantages of everything and kind of, you know, find your way to, um, you know, that vector makes sense for the users and what eventually they're looking for. So I think without domain expertise, and we see that, by the way, even in our go to marketing, right? Without understanding exactly how to sell the solution, what the users actually, you know, is already using, because you're always gonna find incumbent in something else, right?

It's not a new category. It's really hard to operate. So, um, yeah, that's kind of how we see it, and that's kind of how we hire people as well to make sure that they know, they know where they're entering.

Sure. Right. So if I came up to you, you know, um, what's ground cover?

Well, we talked about it a little bit. It obviously is a player in observability and, and so forth, but, you know, level set for our audience, how would you describe ground cover? So I think ground cover is, um, is an observability platform trying to, uh, innovate where there's a lot of pains right now in the observability market.

Eventually, the, you know, the observability domain is pretty well defined when it comes to value, right? You know, traces, logs, metrics, how to create experiences from these values and stuff like that is, is already predefined for a while. Users know what they want, uh, and yet there's so much pain around getting, getting these types of data and, you know, orchestrating in them in a way that makes sense.

And also, um, the, the cost are so associated with that, a lot of different, you know, aspects of that. So Raco is built for ative environments and basically, uh, we challenge other, um, you know, solutions in the market, other observability platform, uh, by using eBPF, which is, um, an advanced technology that allows us to collect data like tracing, like application metrics and stuff that are really hard to get with no hard work for the developers on, on what the one hand, on the other hand, um, completely shattering to pieces, the pricing models that are currently, you know, prominent in the market that make sure that we're aligned with the customer so they can get all the value that they need, like all the logs they need, all the metrics they need, all the traces they need super fast, but not pay an unreasonable exponential price for that and eventually cut the observability sourcing down to just, you know, maintain a budget. Uh, we do that with technology all across the stack.

It's not a different pricing model, but, you know, completely different approach, um, storing the data on the customer premises, collecting it differently. So that's kind of what we're about. Um, making sure the basically users get the most out of what they know they want, right?

They know exactly what they wanna get from an a PM from a log management solution, yet they're kind of handicapped, uh, today for different reasons. So, um, we're kind of eliminating most of the trade offs that people meet, uh, today in, in this, you know, high scale environment. So, Shaha, I want you to imagine someone's watching this out there.

They said, wow, that sounds great. I could probably, I'd like to find out why I'd like, I could probably use something like this. Give them their path.

How, how should they engage with ground cover? So Eeb PF as a technology is the technology that installs super quickly. That's one of the most interesting parts about ground cover.

Uh, it's a zero time to value installing it on a huge environment in, in your cloud will take 60 seconds and therefore, you know, the entire product strategy is built around that. We have a free tier, we have a live demo. You can just install the product, you know, without talking to a human if you don't wanna, right?

Uh, that's part of the, the acknowledgement of our users is not always wanting to talk to a sales person. So you can definitely try it out without us. If you, if you do want, you can engage with us in any one of the multiple ways, including our Slack community and stuff like that, but try out the product, see the power of VPF, see how quickly it can collect data like tracing and a PM value without you doing anything.

I think that's the best way to kind of, you know, start the journey, uh, rather than us convincing you that it's amazing. Uh, we, we encourage all the users to try it out, do Stuff. Yeah.

What's the website for ground cover? com, as simple as that. You gotta love it, right?

Because I, I think you're right. I've been been in tech a very long time. com, et cetera, for 10 plus years.

People like to just get their hands on the product, then see if it works for them, and if they can figure it out, great. If they gotta read the, you know, read the manual as they say RFML, uh, not great always, but, you know, talking to a salesperson, that's the, that's the, that's, that's worse than reading the manual. Um, exactly.

So it, it's an important thing. Now, as you mentioned, the observability market has a lot of noise going on, and it has for a long time, right? And part of it is, is because I think a lot of, um, legacy vendors, you know, did the old Microsoft embrace and extend, if you will, uh, observability to have it include things that maybe originally weren't part of that observability mission or looking at it, you know, using some legacy tools or what they thought of what observability could, should, would be.

But certainly OpenTelemetry has become the kind of backbone, if you will, of, of so much, so many of these observability solutions. I believe Observ, the OpenTelemetry is the second, think of the second or the third largest, uh, CNCF project behind Kubernetes, I think. Um, and, and, you know, look, there's a lot, there's a lot of commercial solutions.

You look under the covers, it's basically all open to telemetry now, ground covers claiming that it, you guys can help eliminate some of the blind spots in OpenTelemetry. Well, before we eliminate 'em, I guess we gotta define them, right? So let's talk about blind spots in OpenTelemetry and how you're eliminating them.

So I, I think the, uh, the idea behind ground covering the, the idea behind EVPF or observability, EVPF is basically just to kind of, you know, entertain our, our listeners for a second is basically, um, a technology built or baked into the Linux car, basically allows us as an absorbability platform, um, to, uh, basically load the simple logic into your, you know, host into your Linux host in, you know, Kubernetes or whatever environment, and basically see what your containerized environments are doing without being part of your code, without requiring your developers to integrate an SDK like OpenTelemetry or do the hard work of instrumenting code for us or the observability vendor. Um, there's a few advantages in that one, you know, time to value is astonishing fast, right? You just deploy an agent on the infrastructure, that's it, right?

No more planning cycles with the r and d to deploy that SDK or work for the observability vendor. And in a hybrid environment, you know, with 10 different code languages, D can cut dramatically, you know, that barrier of just entering into Open Telemetry or an A PM. And the second is that, that in some cases you don't, you can't instrument everything, right?

We, uh, encounter big companies where there's a legacy code they don't wanna instrument, or there's a third party running in their cluster, which isn't their code, right? It's N Engine X, or you know, it's a proxy for something, you can't instrument it. So you will never get the telemetry value from that specific vendor, um, and kind of as you imagine it, right?

So in some cases we see, um, the OpenTelemetry, uh, dream being, uh, you know, further from reality for more customers, uh, that we encounter. So, you know, the series in place, and we totally believe in the standardization of OpenTelemetry and what it brings to the table for perspective of, you know, bringing everybody to the same field. But most companies will never finish this journey.

It's just too hard from an engineer perspective, and in most cases, it's not exactly their day-to-day reality. Eeb PF is kind of a shortcut for that. Everything you have an instrumented all of these blind spots that you didn't get to do or didn't or not able to do fully covered, right?

Right off the bat. And ground cover also announces this week our ability to basically, uh, enhance open demetry where it does exist. So, for example, with EVPS, which is a very low level tech, we can, for example, view the payloads of your requests.

You can see a trace from side to side instrument that would, the actual payload of the API both receive the response from the server side. So, which doesn't exist in OpenTelemetry. So if you have done the work, right, uh, we will be able to enhance your OpenTelemetry tracs with eBPF grade data and kind of bring even more value where you have an instrumented, we can just, you know, basically light up that environment because otherwise you'd be in complete darkness.

So we use EVP for these two metes, you know, places where you're having instrument open phe and places where you did, and trying to bring you as much possible value out of the box without extra work on top of your OpenTelemetry stack. So definitely trying to, you know, to work our way with OPT telemetry, but also make sure that, uh, EEPF uh, paved the way for most of the customers that don't exactly have it, how it makes sense. Exactly.

Excellent. Um, donga, let me ask you just a question between you and I forget everyone else watching. Do you worry about OpenTelemetry?

It, it's gotten so big and it's, look, CNCF does a great job, but they also have 180 projects plus that they manage. It seems like there's so many chips and, and it's good 'cause ground cover has, look, if you don't use OpenTelemetry, they, you can still help, right? Um, but you know, and we hear so much in the open source market about license changes and people, you know, all of these things, you, you worry about that as the CEO of ground cover, sort of what, what's the future of OpenTelemetry?

Um, as, as opposed to what we're doing or that kind of, uh, the question? Well, I mean, I think you have, you clearly have option B and option A, right? With or without Yeah, right.

To quote a U2 saw. I, I think, I think an OpenTelemetry is just a blend of things, right? Uh, it's, uh, on one hand it, the, it's the standard of how you should treat OpenTelemetry.

I how should treat telemetry for observability and how you should measure it, which is great. And I think, you know, eventually bringing all the minds together, and even the distributors, right? Splunk and Datadog and New Relic and everybody to, to the table standardizing stuff.

So you're not vendor locked. That's a great thing. Um, the, the, the problem behind it is, as an engineering product, right?

Eventually requiring me to instrument code, um, it kind of wears out really quickly because it used to work nicely in those homogeneous Java dotnet environments. You know, back in the day when Dynatrace could have done it and Datadog could have done it, and al also OpenTelemetry, and now with, you know, Golang and Rust and like a million different languages running all, all at once on, on my production, it require, you know, the file 30% requires, you know, 85% of the work. Um, so I think it's definitely an important standard and it's great the communities that, you know, continues to build it.

I think it will have to be enhanced by things like EVPS to make sure that customers can actually get value and not just dream about the value. Um, and I think that's exactly where ground cover is PO positioned, we're integrated with OpenTelemetry. We totally believe in the project.

We're gonna contribute as much as possible back to the community from that aspect, but we wanna make sure that our users aren't stuck in, you know, uh, wishing OpenTelemetry would've been instrumented all across their code and have some kind of shortcuts for tracing, which is basically what we do with eBPF. So if you've done, if you've done the work, we're gonna make sure we, you know, we're full on OpenTelemetry back and for you. But if you haven't, then you're gotta get a lot of value that just, you know, other vendors will just, you know, keep you kind of dead in the water waiting for that instrumentation to happen, or, you know, for your developers to make the move of instrumenting optometry.

So I think it's just, you know, a bridge, uh, that, uh, even, you know, helps opt telemetry, bring value. 'cause it doesn't have to be a blocker, right? There's ways to kind of go halfway with eBPF, continue to instrumentation and go from there.

That's how I see it for the long run. I love it. Shaw.

We're about outta time, man. But you know what, excellent. I, I really think from people who are already kind of OpenTelemetry savvy, I think we had a great discussion of like, what's the current state of things and things they could do.

And for those who are aren't, I think we've also given them a great base of, of, you know, knowledge that they could build on. com. Um, will you be at the Cube Con in London?

Will you guys be there? We'll, we'll definitely be there. Yeah.

We, we'll be doing video maybe. I'll see you there, man. Great.

Happy to see you again. Thank you so much for having me. All right, the pleasure.

Yaha Azule, CE Azule, CEO ground cover. com. Check it out.

If you're, if observability is something you're interested in with OpenTelemetry without good company, check them out. We are on Techstrong tv. We're gonna be back in a little bit.