Pablo Stern, ServiceNow | ServiceNow Knowledge 23

This is Techstrong tv. Hello and welcome back to ServiceNow Knowledge 2023. And we're here with Pablo Stern and we're talking about all things observability.

Pablo, welcome the show. Thanks for having me. Mike.

You Guys extended the reach of the platform at this show and kind of rolled out some stuff where you're reaching up to the cloud and other places. Walk us through what the platform currently does and what the ultimate aim game is gonna be. Sure.

So we announced Cloud Observability and Mike, as you remember a couple years ago, we acquired Lights Up and part of the theory around Lights Up and ServiceNow was how do we bring what we observe to the actions that you want to go and take as an organization? And if you look at the state of the art of a lot of the platforms out there, you have observability, but it ends there and the action path isn't one that's connected. And so what we're announcing today is how we can bring from a cloud perspective that observability through the workflow platform at ServiceNow to go and drive outcomes.

So as an example, when you observe a a problem in your digital products, you can now not only quickly diagnose and find the root cause, but then you can go and dispatch and get the right SRE team, the right personas to go and troubleshoot that and then drive that outcome, which would be solving a problem in that digital estate. How do I instrument all those workflows? Because historically at least observability required some sort of an agent to act on.

So how are we gonna go and execute on that? Yeah, so I'd tell you a couple of things. One, in the cloud world with open telemetry, one of the things that's happening is the true Democrat democratization of how you can go and find in those environments the sensors of where you're having issues and problems.

And with our cloud observability, we connect directly into that world and we can drive a few outcomes from that. It starts with just giving you visibility into those estates. They're massive, they're very distributed.

And you need to understand not only what exists there, but what are the applications and services that you have and who are they tied to so that there's an issue you can drive resolution. And so that's the first step. And the second one is once you have that, then you can go really drive introspection, whether it's at a trace level from the metrics that you get or logs in those environments to then understand where the problem is, what the probable root cause is.

And then the workflow piece, Mike, that you mentioned is where ServiceNow has strength. That's the power of what our platform can do. We can connect those workflows to drive outcomes, whether they're self-service or to the right person at the right time to go drive re resolution on those problems.

Some people will say we've always had some form of monitoring. What is the difference between what we're calling observability and what we used to call monitoring? Yeah, so the perspective within ServiceNow is it starts with the foundation.

And what we've seen with our customers that foundation is the, the marriage of service and operations. So the construct is that you have service management practices, which is how you're driving issues like a major incident or uh, incident response. And then you have your operating estate, which is, you know, the hundreds, thousands, or tens or thousands of different applications or microservices that you have there.

And so the starting point for us was combining those two in this construct of service operations, which you could do on the ServiceNow platform. And by doing that, we connect the people to the machine. That's the first step and we can drive workflows around that.

Now, if you extend that and you go into your cloud environments with observability, we take it one step further because not only we connect to the people and machine, we can actually go and introspect what's happening in those highly distributed environments to find out those needles in the haystack. So when you identify them, not only do you observe, but then you can act. A lot of people would also say, it's wonderful that I can observe, but I have no idea what question to ask in the first place.

So I don't know what to interrogate. So how do we give people the guidance to go ask the right question? Yeah, oftentimes when there's a problem, the questions you're asking is what changed?

And this is another place where I think ServiceNow as a platform helps bridge that gap because not only do we help you observe what happened, what's happening in your operating state, we actually have a view of what changed in those environments because we track and we capture all the changes that go in through our system of record and our service graph. And so by doing that, we can go and then really figure out from that, observe from that observation, what was that changed, the cause of problem, which is most of the time where those problems emanated from. And then drive a workflow that gives people the right level of, of insight into the changes that could have been the problems so they can go and diagnose and resolve quickly.

Observability is a term that is heavily rooted in DevOps. Are we starting to see, um, DevOps and ITSM and workflow management all kind of converge? A hundred percent.

I think that if you look at the state of the world and how our customers are trying to evolve their digital products and services, you can't live in a world with manual cabs, manual processes, broken games of telephone tossing over the fence from a product team to an engineering team to an ops team. And so we've been very focused on the product side to connect those, to connect into the DevOps pipelines that you have and help our customers shift left, move more changes to automated workflows, understand what your security posture is as you're putting stuff into production, understand the estate as it goes into production so that when things are there you're not as reactive, you're actually proactive and that enables teams to drive resolutions to issues much faster. What is the connection between observability and the collection of all the data that we need to observe and the application of AI in the future?

I mean, it seems like the two are joined at the hip. Yes, absolutely. And you know, AI is definitely the topic du shore as it relates to, you know, everything that's going on from a generative AI perspective.

But if, if you look at ServiceNow, we've been focused over the last four or five years in how do we drive some of the machine learning that's gonna get you to some of those root causes more quickly? And it starts with understanding that estate being able to quickly identify and troubleshoot and correlate where a lot of those issues are. And that truly is that world of AI operations that gets you to some of that true introspection.

And then with the power of generative ai, we can actually help get summarization of issues, potential root causes, potential solves around those root causes by looking across an A state and trying to understand not only what is the diagnostic of the problem, but how has it been solved in the past, what are those root causes and then how can you go and take action on it? So I think we've done a, we've made a tremendous amount of progress on it and I think that generative AI is really helping open the doors to a whole new world of how you can get to those outcomes much faster. How smart will all this get?

Can I just walk into my office one day and verbally express, tell me the three things that are likely to get me fired and what I should do about it? I Don't, I don't know Mike if you're gonna be able to do that, but I do think that from a, uh, an AI perspective, the way that we see it at ServiceNow is, you know, now we'll assist you to basically be able to drive those outcomes. So we basically end up being enabler to a lot of the different teams.

Whether you're a site reliability engineer, you're an IT operations team, you're a service desk agent with now assist, we will help you and we will give you the knowledge that you need. And so you can almost think that as an amplification of what you're doing and you're bringing the power of the machine in AI to help you do your job more effectively, more efficiently. One of the challenges that people encounter when they first go down the observability path, what are you seeing from folks and what would you recommend as, you know, what are the steps to get there?

Yeah, so I think starting always with the end in mind is how I think of things. That's what are the outcomes you're trying to drive? And from an observability perspective, like if you're really thinking of the world that you have and delivering amazing customer or employee experiences, what are the things that you need to be able to do that and how do you interconnect that outcome to the systems that you have?

And with many customers, you may have multitude of systems and tools that are helping you get to some of those outcomes. And from a ServiceNow perspective, the way the workplace we've been very focused is to bring that all together to give you that central control tower so that as you're driving those outcomes, you can get to them faster. And you don't have to swivel chair as much as you have been in the past.

We have had islands of automation for years and islands of observability. So, um, do these, the walls between these things need to come down Cause it seems like everything is kind of tangentially, at least related to each other and you can pull one string, all kinds of things happen somewhere else. Yeah, I think if you look forward in the world, like there are definitely gonna be multiple different places where you're gonna have sensors and environments and try to understand what's happening.

And there is gonna be a need to drive from an AOPs perspective, a correlation and a, uh, normalization of that back to a central system of record. And the way that we've been focused on this, Mike has been we will go and drive from a full closed loop perspective, we can get you from that ob ob observation right to the outcome from a ServiceNow perspective. But we also recognize that there's an interconnected world that is more heterogeneous and we bring those in and that's where our AIOps and service operations comes in to make sure that we are correlating everything that you're seeing to drive those outcomes.

So I think in the end there are definitely some efficiencies and if you talk to most customers, they'll say, look, we know we can get to like a more rationalized state from what we're doing from an observability perspective. And I a hundred percent agree that that's possible. And I think that automation from a ServiceNow platform perspective is a way that you can drive those outcomes to not only help you get more efficient, but ultimately get to the outcome that you're really driving for, which is faster resolution on issues and more and better experiences for the folks that are ultimately ingesting and using those services.

I know we have called it IT service management over the years, but it was always really IT management. Mm-hmm. Are we shifting more towards where it's really a service management construct rather than just trying to figure out which applications happen to be available?

The way I think about that evolution, and I've seen this, uh, in a lot of conversations with customers is that that world of the, the evolution of the service from how we, how you're managing the full life cycle of what you have in your operating estate is actually a convergence of that service estate and the operating estate. So I think that the two B state for a lot of what you have in operations is actually a construct that the telcos have had for a long time, which is that notion of service operations. And I think that is the foundation for what you do and how you manage the services that you have that are out in production that are being used by your employees or your customers.

So am I gonna see people get certified in service operations? Is that the next thing? I hope so.

I really hope so. I do think that, you know, the construct resonates tremendously from, from a customer perspective. I dunno if there'll be like certifications per se, but what I do think you're gonna see is more and more adhering to that construct, the move from like, you know, your network operation centers that we had in the nineties to your command centers, that teams had to more of the telco model, that service operation center.

And I think that is the evolution of the state. And again, it's about how do customers get to those outcomes and how do they interconnect the people in the machine so that you can go right from that observation to the outcome that you're looking for, the resolution of that issue back to the root cause as quickly as possible. Do we need to change the way we're organized to achieve that goal?

Because, um, today we have people who own the applications and then we have people who run the infrastructure. And yet if it's an end to end service that I'm trying to create, much like a microservice, do I need a team that says, this is the thing you want? Yeah, I think that there is an evolution of like the ownership model and you know, whether it's the team that owns the infrastructure or that owns the reliability of the site all the way to the application layer, we see this across our customer base.

It is evolving, I think like there probably isn't just one model that rules 'em all. And so from a, from a product technology perspective, ServiceNow is very much focused on making sure that we can enable multiple different models that our customers have. I do think that over time you're gonna continue to see more and more is owned directly by the team that's actually building that solution to drive the full closed loop because they, they ultimately have most of the knowledge around those products that are gonna be able to drive resolution to those issues.

So you can almost think of that as like moving up the stack to drive the resolution. The reality ends up being it's probably not the final destination because you still need to have certain expertise that are required at different layers of that infrastructure. So again, I think we'll see that evolve.

I do think that, you know, the solutions that are out there need to be able to support multiple different models and they have to have that flexibility built in. People have been talking about the divide between IT and the rest of the business, but you know, increasingly the business is it and the two are converged. Is much of what we're talking about here a cultural issue as much as it is a technical issue?

Yeah, the the reality is that, you know, from a, if you think about it from like the outcomes you're trying to deliver, a lot of technology teams and IT teams are very much focused on an outcome that is supporting either a business outcome, an employee outcome, or others. And that's how they think of it, right? They think they think about the outcome, and then under the covers, you need technology to be able to deliver that outcome and you have to think about the processes and workflows that you have that drive them.

And so I do think that over time what you end up seeing is it is about driving more of a business lens to the outcomes that you're delivering and making sure that the technology that you have is supporting that end outcome. And one of the things that we've seen with a lot of our customers is as more and more moving towards digital, what we're initially employ back office outcomes that they were focused on from an IT lens are now also customer facing outcomes that are driving the business. And so it's not only about your employee and employee experience, but also about your customer experience.

Are we underestimating the complexity of the service? Because all these things seem to have dependencies on something else and some of the dependencies are hidden, some of them are third party things across an api. How do we get that observability at that level of scale?

Yeah, the, you know, the digital estates are continuing to increase for our customers and the level of complexity. And we see this, you know, in like these cloud environments with microservices, you see it as you're tying to different services where you're calling from an API perspective. And so one of the things we often talk about is the first step is really driving that visibility into that estate and making sure that you can quickly and accurately pull in what you have so that you understand how things are interconnected and interrelated, and then being able to drive the observability into that environment so they can go drive outcomes.

So it is becoming a problem that is harder to solve, and this is fundamentally why we think of that service and operation convergence being so critical because in the world of old, you are doing it just with people and tribal knowledge and that just doesn't scale. And so if you can bring that operating state, that system of record view into that world, you can actually start driving these outcomes more effectively and efficiently and move away from that tribal knowledge or potentially that broken and disconnected system level view. Do you think we'll get better at finding the gremlins?

You know, that one thing that intermittently happens every three weeks and no one knows why and we just accepted and lived with it all these years. So will we get better at hunting down the gremlins and the root cause? Yeah, I think, I think we will, I think we will find out where the gremlins are.

I think we'll help identify where you have single points of failure. So not only where the gremlins are, but where the potential gremlins are gonna go hide next to help you understand those environments, which can then feed back into how you think about the value stream of how you architect these services for reliability. So it ends up being a virtuous cycle that you can actually create and enable.

What's your best advice then to customers? Let's say we make you observability king for a day. What's that one thing you would impart upon your subjects?

It's a, it's a funny construct, Mike. Um, I would say if I think about the world from an, in terms of like observability, I think you have to really think of the end-to-end outcome that you're delivering. And so observability is gonna be key and it's gonna be ever more critical in these like massively distributed states where you no longer have a three-tiered architecture that like, you know, with tribal knowledge, three or four people can actually get to root cause quickly, but you only have hundreds or thousands of those services that are interconnected.

And so you need to one, drive visibility into those services, make sure you understand them, have a way to then diagnose and quickly observe what's happening in those environments. And then the, the critical thing is make sure that you can go from a observation to action. Like what is the way that you're, if you observe something, how are you acting on it?

How are you getting the right teams to go drive the resolution as quickly as possible because you need the full closed loop. All right folks, you heard it here. Observability is the difference between just seeing something and actually knowing what it is.

Pablo, thanks for being on the show. Thanks Michael. All Right folks, that's a wrap for today.

We'll be back here tomorrow. But thanks for spending some time with us and we look forward to doing it all again. Take care.