Tag: microservices

Mitch Ashley: Hi, everyone, this is Mitch Ashley with devopsy.kinsta.cloud and you’re listening to another DevOps Chat podcast. Today, I’m joined by Rick Fitz, SVP and GM of the IT Markets Group at Splunk. And tailoring onto some recent news, the topic is about Splunk’s intention to acquire Omnition. They announced that on September 4th. Rick, welcome to DevOps Chat.

Rick Fitz: Thanks, Mitch. Happy to be here.

Ashley: Great to have you on. Would you start just by introducing yourself, tell us a little bit about what you do? And I’m sure most people know Splunk, but maybe a sentence or two about Splunk.

Fitz:  Sure. A little bit about me, I’ve been building commercial software my entire career, and started out as a developer working for companies that were development oriented, and then the last 15 or 20 years, I’ve been working on just developing software for IT operations and application development teams in a commercial fashion.

And in my new role now, I’m here as a General Manager, so I’m an executive here at Splunk, responsible for our strategy and vision of what we’re doing in the company. For those who don’t know Splunk—shame on you, darn it.

Ashley: [Laughter]

Fitz:  [Laughter] But let me tell you a little bit about the company. We are the company that actually built and invented, if you will, the ability to index logs at scale. And it made it really easy for the IT practitioners of the world to search that data and find, through an investigative lens, find the actual answer to those murky problems that are always out there. And so, the company is roughly 5,000 employees now, and growing very rapidly. So, we’re headquartered in California.

Ashley: Well, just a little bit of trivia. I can’t claim credit for this, but I came across Splunk in probably 2006 or so at, I think it was an RSA conference. Tiny little booth, great ideas. Like, this is a cool idea. And of course, the company’s done really well, and I later became a customer of Splunk. So, congratulations on all the success and glad to see that you’re there, too, of course. Well, let’s talk—

Fitz: Well, I thank you for being a customer.

Ashley: Oh, absolutely, absolutely. [Laughter] So, let’s talk about Omnition. A lot of people may not know about them. Maybe they do, if they’re open in the—or involved in the OpenCensus, OpenTracing, OpenTelemetry projects, if I can say that right.

Fitz: Mm-hmm.

Ashley: They may know them from that, but they’re in ________ so there hasn’t been too much about Omnition. Tell us about that company.

Fitz: Yeah. The—first, I’ll back up a little bit and kind of let you know why we even picked up a company called Omnition that no one really knows about. [Laughter]

Ashley: Sure, that’s a good question, too.

Fitz: The reality is that, in these large, cloud native worlds that a lot of us are starting to develop in, all kinds of things are changing. The things that we build on, the infrastructure we build on, the compute surface, the storage surface, there’s all these things that are now services that are available to us to leverage as application developers.

And the other thing that’s changing pretty dramatically is just the way we actually build applications. The way we try to attempt to continuously integrate them, continuously deliver them. And perhaps one of the most profound things that’s occurred is that developers are now being asked to kinda own their code and they’re being asked to be on call or being asked to resolve issues.

And so, if you will, they’re starting to become, as your website suggests, DevOps professionals.

Ashley: A la DevOps.

Fitz: Yeah. [Laughter] These trends start to kind of cause a paradigm shift—well, they have, actually, created a paradigm shift in the way you monitor. So much so that, probably over the last couple of years, DevOps practitioners have been talking about monitoring as—they’ve labeled it as a new world called observability.

Ashley: Mm-hmm.

Fitz: You know, I’ve been here around enough to know it’s simply monitoring. But I think what they’re trying to point out is that you need to observe the behavior of the data from the systems in order to make your lives easier. And that data comes in the form of logs, which is what Splunk has been known for a long time. It comes in the form of digital exhaust coming from monitoring tools in the form of metrics, which is really important.

And then the last thing—and if you’re an application developer, I think we can all kinda relate to this—is the traces, the traces that actually come from the application code itself being executed are really great uses of good information.

But all three of those things combined, the logs tell you kinda the root cause, the metrics tell you kind of what’s going on—is there a measurement that’s out of whack? Do I need to do something proactively to cause something to remove an error?

And then the traces are really good for that logical layer—that logical layer telling you where to look. Do I have a backing problem? Do I have a code problem? Is it my problem, or is it a front end problem? And that kind of isolation is really important to trace it.

So, that leads me to answer your question is that, you know, Omnition. Omnition was actually founded in 2018. The team is very small. They were in stealth mode, meaning they were building an application that was focused on tracing, and very specifically, they had a couple tenets in mind when they wanted to do this. They wanted to leverage open source. So, you mentioned that they’re big open source contributors, specifically to the OpenTelemetry project.

And the other thing that they wanted to do is, they wanted to build this logical topology model that would allow you as a practitioner to isolate where a problem resides. And that was a pretty—it is a pretty novel way of looking at this particular problem. And it’s important, because having a thousand traces isn’t terribly intuitive or useful for a developer.

Ashley: [Laughter]

Fitz: But being able to isolate that it’s a front end problem so I can go talk to the guys that are in charge of that—or maybe it’s my logic that’s causing it, but maybe it’s a network issue, firewall issue or something like that. Or maybe it’s outside my firewall and I’m dealing with some latency issues from the world around us.

Ashley: Mm-hmm.

Fitz: But you need to be able to isolate that problem so you can solve it. So, long winded answer is, these guys, this team really figured out how to build this logical topology sitting on top of this trace data, leveraging open source—so, pretty darn powerful.

Ashley: What I was really excited about with this announcement is, one of the things that I talk with companies about and work with them oftentimes is what DevOps, in this fundamental way that we’re creating more cloud native applications, how this really changed the way organizations work as well as how we create software, which I know you speak a lot about.

And in some ways, what we’re doing is through this kinda granularization of functionality into containers, microservices, mesh—service mesh—it’s very interesting, because you can do lots of great things with it. But it makes the complexity of operating and monitoring—as you say, observability—even more difficult. So, that’s created this whole new cottage area, if you will, which is what Omnition went about, and actually some of the, they’ve had some developers who came over from Lyft, I think, went over to Omnition to help solve this problem.

So, it’s fascinating to me to see Splunk bring this, this quickly, into your kinda product portfolio of companies, or at least that’s what we expect to happen with the acquisition, that you’re jumping right into service mesh, microservices, and observability, the monitoring of that. So, it seems like it’s a natural fit—maybe, you know, sooner than I expected it to happen, actually.

Fitz: Here’s what’s interesting. I mean, we have the luxury of a lot of—you know, being around for a while and talking to a lot of our large enterprise customers that are going through this digital experience. You know, they’ve done their lift and shift, they’ve moved to kinda the lift, shift, and augment, and now they’re building a whole new series of cloud native digital services. So, they’re building in this new cloud native way. And when you start to talk to them, they recognize the problems, and they realize there has to be a way of doing this.

And what we’ve found was, a number of them were trying to build it on their own, or they cobbled together a prototype that actually would kind of get ‘em there, give them some visibility. And then having a lot of those customers kind of advise us—because they’re very large customers at Splunk because of the fact that logs are a critical component of this strategy. But they were building the rest on the side, and in some cases, you know, using other companies to try to do it.

And we just quickly realized that this is an opportunity for some of our largest customers, for us to come in and help them with this very critical issue and build alongside them as we move forward.

So, yeah, we recognize that certainly the tracing part of this world is pretty new, but we think it’s the right approach, it’s the right answer to a long-standing problem, and I think the application architectures are at a stage in this cloud native world where this is perhaps the only way that this is gonna be accomplished. So, that’s why it was so strategic for us to do this.

Ashley: You know, like, observability kind of being replaced for the term monitoring, tracing is also used—you used the word “also”—as kind of a new term in this service mesh, microservice world. Does it have a fundamentally different meaning to you as to what log analysis and things like that—is it substantially different or is it just an extension of what we’ve been doing in a non-container, non-microservice world?

Fitz: Well, the only analogy is, it tends to be text, [Laughter] and you have to put it somewhere. And actually, some of our customers were putting it in the log and then extracting it, but that’s useful if you’re looking for one specific trace, but if you can imagine some of these applications are throwing up hundreds of thousands of traces a second, and sifting through that, while you can search through it, it’s great, but you really do need some analytics on top of it to organize it in a way that creates useful information. And that, I think, is again what Omnition has done.

You know, answering the question, “Is this different from the old approach?” You know, I had the opportunity to work with an early Wily team. They had been acquired by CA Technologies, and I got to work with the team, which is an APM company. And the team actually invented bicode instrumentation, which is a way of doing instrumentation of the application on the fly. And it was very novel, because at the time, the operations teams were responsible for monitoring and the developers weren’t. So, the only way they could get their developer team to get visibility into the application—because they couldn’t tell their developers to actually put stuff into the logs.

Ashley: Mm-hmm.

Fitz: Well, they could ask. [Laughter]

Ashley: [Laughter]

Fitz: But they just couldn’t demand.

Ashley: They might have ________.

Fitz: Exactly. [Laughter] You know, us developers, how we are. And so the funny, interesting thing about that is, we tended to sell to the operations teams, so they could get that visibility into their application. In the open tracing world, because the onus, and it gets back to this DevOps ________, the onus is actually upon either an SRE function or, if you will, a DevOps function to establish the standards and policies. Those policies have been enforced by that team, and therefore, the developers, before they can actually bring product into production or bring code into production, there’s ________. And now, all of a sudden, you’re getting a lot more visibility and a lot more fidelity inside the code by that enforcement.

And then, of course, open tracing is now on to its own standard, so you’ve got a standard to kinda drive. Before, there was no real standard, so you were just putting stuff into a log. And therefore, it’s hard for a vendor to look at it and develop an application to give you visibility into it.

So, that’s definitely some of the big shifts that have caused this new paradigm to actually emerge, which is pretty awesome.

Ashley: I think I saw in a blog post either on your site or Omnition’s that their investment and contributions to OpenTracing is something you’re likely to continue and support and want that to continue, is that correct?

Fitz: Oh, absolutely. Absolutely. We’ve actually been, at Splunk, probably for the last four or five years, pretty definitely contributing to open source more, leveraging open source. We’ve been kind of on that bandwagon for quite some time. But this one, it makes a ton of sense. Both the OpenCensus and OpenTracing have now converged into OpenTelemetry. It makes a lot of sense. And I think that this is definitely where the world is—where it’s headed. I think it’s long overdue, actually. [Laughter] I’m glad to see these things kinda moving forward.

Ashley: Given the players involved, I think if I recall right, OpenCensus was started by Google, but quickly followed by Microsoft and Dynatrace and Omnition—

Fitz: Yeah.

Ashley: – I think was part of this effort, too. So, it’s got some important names behind it.

Fitz: Yeah, and most of the monitoring companies out there that are touching on applications are participating. So, I firmly believe it will take on as a standard. I think that’s pretty clear.

Ashley: As you were considering an acquisition like Omnition, what were you hearing from your customers that led you to say, “We need to go find a company” or, “We found somebody like this. This would be a great fit for some of the problems and needs requirements that’s awesome we’re hearing from customers”? What was it that’s behind that, what you’re hearing from customers?

Fitz: Well, there’s the big part of kind of the equation was this idea that we can use OpenTelemetry and spit this data into a log. But that’s good. I mean, that’s a great first step. I can search for something and that’s a fantastic, great step.

 But very quickly, the customers that we were dealing with at scale, they were saying, “You know, what’s interesting is that all this data—I have the data, but I can’t find any useful information out of the data, you know? I’m struggling with that.” And so, you know, we were in there building, on top of Splunk, the ability to model it and do different things. And we realized there’s the answer. The answer is to develop some form of analytics on top of this big, rich set of data that provides you really useful insight.

And then, from my APM experience and FHIR experience, you know, the whole idea was having kind of a logical topology that represents the application, and I mentioned some of those words meaning back ends and front ends on the application itself. So, once you have that meta model in your mind and can decorate it with trace information, it becomes incredibly powerful. And that’s—when I broach the subject with our customer and said, “Is that really what you’re looking for?” they said yes.

And so, we’re surveying the market and seeing is there anything out there, and then Omnition popped up, and it was like—wow, it’s a good day. [Laughter] It’s a perfect day for us, because we would’ve had to build this, and I think this team, with the expertise they have and the smarts they have is gonna be a welcome addition to the team and help us do it for us.

Ashley: They do. They do have some great people. Now, Omnition, as I understand it, they describe themselves as a SaaS company. I assume this is a SaaS service that they’ve been building. Is it something new for Splunk in terms of a SaaS offering, or does this fit in well with other things that you’re doing?

Fitz: No, it fits in very well. You know, SaaS grows from a business perspective as a very high growth business for us. We tend to, because data has gravity, [Laughter] we tend to have our software where the data resides. But more of that data is starting to reside in the cloud, so you know you’re starting to see people, as I mentioned, lift and shift or lift and shift and augment or building kind of in a cloud native way.

So, we’re definitely—we’re part of that trend as well, and we think it’s the smart thing for all of us to get there. So, yeah, we are definitely gonna be SaaS on this particular offer, for sure.

Ashley: Well, that makes sense, too, doing cloud native if you’re doing microservices, service mesh—more than likely, that’s gonna be stuff in the cloud. So, having Omnition there as well, that seems like the place to be, the right place to be.

Fitz: Yeah. You know, it’s interesting, I was telling some sales reps yesterday, we were talking and I said, “You know, I contrast the conversation that I had with somebody who owns the data center with somebody who’s working in a cloud native way, and the vocabulary is completely different.” [Laughter] You know?

Ashley: [Laughter]

Fitz: Nobody says on premise or cloud—they just don’t think that way. They just assume that it is all cloud, you know? Whereas if you talk to somebody that’s still in a data center, they wanna know where the data is, where do I stack it, where do I rack it, where do I plug it in?

And so, there’s definitely a difference in that cloud native world, the way people think.

Ashley: Well, I’ve often had to believe, maybe learned over time, is that you can tell who drew the picture of a diagram on the whiteboard. Whatever is at the center is the data—if it’s the data center people who drew it, if it’s the cloud, it’s the cloud people.

Fitz: Yeah.

Ashley: But it’s interesting the transition we’re making with multiple companies investing in cloud management platforms, even down to storage providers and VMware and others. You know, I think eventually we’ll stop talking about hybrid. Everything’s a cloud, whether it’s just in your cloud or somebody else’s cloud and it’ll be managed much the same way. That’s where we’re headed, anyway.

Fitz: Yeah, and I kinda share that point of view. I think, organizationally, we have to kind of force some of that standardization of vocabulary just to make our lives work. [Laughter] I think it’s gonna get—

Ashley: Well, it’s also—yeah. [Laughter] I also think you have to do it from a cost standpoint, right? Complexity, managing a physical data center one way and then all these different clouds a different way. So, anyway, a whole ‘nother topic for another time.

Fitz: Yeah.

Ashley: We have just a little bit of time left—is there any message that you wanna make sure that gets out to listeners, to customers, future customers about the path that Omnition sets you down or other things about this pending acquisition that we wanna make sure folks know?

Fitz: Yeah, let me just kinda share a couple things. One is, as I opened with, I think this world of observability is certainly all about having logs as well as metrics as well as traces. Just a couple weeks ago, we actually also announced our intent to acquire SignalFx, so we’re bringing in some of that metric capability. We’ve been known for years to be the log company, and then of course, Omnition is helping us with our tracing capability.

And so, I think the combo is gonna be pretty darn powerful. I’d encourage your listeners, if they get a chance, if they wanna join us in Las Vegas this year at our .conf, which is our conference. We’ll be talking a lot about what we’re doing there and getting everybody up to speed. And if you can’t attend, all that stuff is posted online, too. You can check us out there, as well.

Ashley: Yeah, there are definitely some great videos out there, too. Well, I wish you all the best with the acquisition and future business with Splunk. I just want to say thank you—Splunk is a great product, and when I had the opportunity to introduce it in organizations that I was running, all positives. So, it made a big difference in our world, and I think your move into observability around microservices, et cetera, is fantastic, and we appreciate you taking those steps.

Fitz: Well, thank you. Thank you, Mitch, and thank you to all the listeners for your kind words. I’m sure you’re gonna share with me online later on.

Ashley: Well, thank you, Rick—Rick Fitz, SAP and GM of the IT Markets Group at Splunk for joining us on the podcast today. And, of course, I’d like to thank you, our listeners, for joining us as well. This is Mitch Ashley with devopsy.kinsta.cloud, and you’ve listened to another DevOps Chat. Be careful out there.

— Mitchell Ashley