2023 Observability Forecast Report – Peter Pezaris, New Relic

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

Well, you know, tech Strong TV travels the globe, trying to find guests for our show here today who are important and have important things to say for you. But sometimes we just have to go in our backyard. So I'm really happy to have today one of the few people who actually live near our tech strong headquarters, Peter Azaris.

Peter is Chief Strategy and Design Officer at New Relic. Hey, Peter. Neighbor.

Welcome. How are you? Hey, it's great to be here.

Thanks for having me. Good to see You. Fantastic, Peter.

Um, I mean, besides living close by our office, why don't you, if you don't mind, give people a little bit of your background and your role at New Relic and, and maybe you could weave in for, I think our audience is familiar with New Relic, but there might be some folks who aren't Sure. Sure. Give A little background.

I'm happy To. So I joined New Relic two years ago, a little over two years ago by way of an acquisition. I've been a serial entrepreneur in my past doing tech startups, and our latest company is Code Stream, and it's a developer tool that makes it easier to collaborate as you're writing code with your team.

Mm-hmm. So, new Relic acquired our company two years ago, and I currently serve as our chief strategy and design officer. I sort of wear two hats.

One is head up corporate development, business development and partnerships. And then I also run our design department New. Love It.

New Relic itself is a cloud-based observability vendor. We're actually pioneers in the application performance monitoring space. 14 years ago, we created the first cloud-based a p m service, and now we've grown into more of a traditional full stack observability company where we offer tools to monitor your, uh, a p m, your infrastructure security logs, uh, and we allow you to really get a good handle on your entire production estate.

So you can both monitor what's happening at any given time, but also troubleshoot and fix things when, when they inevitably, uh, when you have a problem in production. Yep. com 2013, we launched first, published 2014 early, uh, new Relic was already one of the foundational players, right?

Right. I mean, they kind of defined the a p m space along with, uh, AppDynamics. Right?

It was AppDynamics or New Relic that that was a p m puppet and Chef and Ansible were, were over here, new Relic AppDynamics there, and I mean, and, and that was kind of the, the layer and then the CloudBees Jenkins, right. With C I C D for all intents and purposes. And GitHub, that was, that was your DevOps tool stack right there on one hand.

Um, now of course there's 800 companies or some more probably, but it was a lot simpler that, but one of the interesting things is we talk about Shift left. Yep. Right.

And, and in a lot of ways New Relic has shifted left too, where, you know, a p m was application performance monitoring post-deployment. Correct. As you said, code Stream is, was a, is a tool developers use to improve their code in real time.

That's right. And you know, very much on the other side of the, of the, uh, you know, shifting left on the other side of the deployment barrier, if you will, membrane. Um, and so, you know, just as DevOps has shifted left, so has New Relic, right.

Put this emphasis on shifting, uh, shifting left. And, and it's also the story of observability, right? Observability, we observability doesn't start a deployment anymore.

Observability starts when you first put pen to paper if, well, no one puts pen to paper, but you know, when you first put code to keyboard or something else to keyboard Yeah. Finger to keyboard. Good.

Yeah. Um, That, that's super consistent with our strategy, you know, and I think a big part of the thesis behind the codes from acquisition is that new relic's mission is to empower engineers to make decisions with data, not opinions every day at every stage of the software development lifecycle. And that speaks to the shift left principle that you just mentioned.

So what we're doing with Code Stream and the integration within New Relic a p m, is we're bringing all the best parts of a p m and we're putting them into your I D E. So now as you're writing code, you see how that code is performing either in pre-production or in production environment, and you can compare. So whether you're doing a code review, whether you're writing brand new code, or you're just, you know, learning more about how your application works, now every engineer on the development team can see those insights because we're in the tool they're already using.

Absolutely. You know, that, that, I, I told you I added a conference last week, and that was a big theme there that tools are moving into the I d e it, it's almost a shift, A more left shift left, yeah. Right.

Where we we're moving into the I D E, because that's where we can really make stuff happen early on, where it's, it's easier to, to influence that stuff. Correct. Um, Peter Observability, look, you know, it's the new black in in, in, in a lot of it.

Whether you're an SS r e or DevOps a developer or what have you, uh, everyone needs observability. We've had a lot of change, not changes, let's call it evolution to observability over the last, let's say 18 months, certainly over the last nine months with generative AI and, and all that this brings to observability. Now, observability, as we were talking off camera observability was actually one of the stacks, if you will, where AI and ML were already firmly entrenched, right?

That that's kind of cornerstone observability stuff, looking for patterns in large amounts of data and, and then making actionable intelligence from it. Generative AI is having a, a profound impact as it is on everything it seems, but really on observability, you guys recently came out with your New Relic 2023 observability forecast. That's Right.

Interested to hear how these kind of mega trends are playing into your forecast, what you see in the market. Yeah. Um, so our observability forecast is the third annual forecast that we've done, and every year the report gets a little bit bigger.

This year we surveyed 1700 IT professionals from 15 countries across the globe and asked a series of questions. Some of them repeat questions from last year and the year before, so we could establish some trends. And some were new questions.

Like you were talking about a brand new, you know, we didn't ask them about generative AI last year. No, Who did. But, but this year we asked some new questions as well.

And some of the insights, I would say the most profound insights, um, were ones that are part of the sort of industry and the usage of observability as a whole. Because although there's tremendous amount of excitement around generative ai, I'm not sure anybody really understands the impact yet. So there is less, uh, we, we receive less data about that.

Now I can share my opinions and, and what my thoughts about generative ai, but I think the data that we got from this report is pretty compelling, you know, and there are a number of very interesting findings that come out of so much data. And this report is published on our website, and you can find it and even download the data if you want to crunch the numbers yourself. Um, we've done that, uh, make open source the data to make it easy for people to understand and get their own insights from the data.

Yeah. You can slice and dice it as, as you want to there. That's fantastic.

Um, of course, new Relic dot com's website, they could download the, the report from there. So Peter, let's, you know, you mentioned some profound insights. Let, let's hear some.

Yeah. Um, so the one, a few numbers really jumped off the page for me. Um, one is among 1700 IT professionals surveyed the median annual ca cost for outages.

75 million, that's Actually, that's higher than I've seen in the past. Yeah, that's a pretty big number. And that's A big, that's a chunk.

And a third of, of the respondents say that outages cost $500,000 an hour or more. Yeah. You know, so my background security, right.

So this is something like the Verizon data breach report has, has been not exactly this metric, but they track the average cost of a breach and so forth. Yeah, exactly. And, and downtime as a result of breaches.

Um, go ahead. One of, One of the things that may not be a surprise, but it's sure feels good as an observability vendor, is that those respondents who have achieved full stack observability, they experience much fewer outages for a much lower cost. So it's 30, actually 37% lower annual outage costs if you have full stack observability.

So Connect the dotted lines for me there, Peter. Why, why, what I why do you think anyway? Of course, you know, Sure.

Um, a lot of companies are working on their observability maturity. So the best practices include monitoring your entire stack. So that includes front end applications, services, the environment, or infrastructure security and logs.

So those are the five elements that we define as full stack observability. So that's sort of table stakes for having a mature observability practice, but there are many other elements. For example, having what customers tell us is they actually want to have a single vendor or the minimum number of vendors when possible.

So if you have too many observability vendors, then you suffer from something called screen swivel, where you've got data silos and you have, for example, your logs data with one vendor and then your a p m with another vendor. And so as you're troubleshooting a problem, each time you think you're getting close to the answer, you have to switch to a different screen, switch to a whole different analysis engine, and you're basically starting from scratch. And that makes a big difference.

You know, when, when production is on fire, rather than grabbing a fire hose, now you're sort of trying to figure out how to log on to some of their vendors' website. So we believe, and our customers are telling us that tool consolidation can be a big factor in getting better outcomes. Yep.

And just, just to kind of put the bow on this, right, so by, by having one vendor or, or fewer amount of vendors, less screen swivel Yep. You, your response to an outage is much more organized and, and effective, thereby lowering the, the cost of the outage and the duration of the outage. That's right.

Yeah. Lemme ask you another, go Ahead. Well, customers are, are, there's three different dimensions to this.

Uh, not only customers say they prefer it a two to one margin by a two to one margin, they have a preference for having fewer tools rather than more tools. Um, the number of tools per customer is coming down. I mentioned, you know, we're asking the same question year after year after year.

So on average, the respondents report having one fewer tool than they had last year. And those that have, uh, fewer tools actually spend less money on observability. So it's, it's their preference, it's cheaper, it's better, and it leads to better outcomes, so you save money.

So it really is, I think, a trend that we're, will accelerate and become a standardization. Like people will say, we're standardizing on New Relic. So let me ask a question of you.

Um, you know, I've been in it a long time, man. Yeah. And, and to me that that issue right there is sort of a pendulum issue, right?

We go from best of breed to one throat to choke. Yeah. Every, I don't know, eight, nine years maybe.

It seems like that pendulum goes, I, I think, and this is just my opinion, and I don't hold myself out, you know, to be any smarter than anyone else out here, but I think we see the pendulum swing to the one throat to choke to the fewer vendors mm-hmm. When technology matures Right. Within a given segment.

And, and so there is no one vendor potentially who is maybe heads and tails above or has some, you know, dynamic new technology that no one else has when all the vendors are sort of here. People tend to go for less vendors, greater stability in vendors, meaning public companies of more size, you know, and, and, and presence. And then when we see sort of a, a new technology sort of, you know, uh, era arises where you start seeing, you know, a a Cambrian explosion of, of, of new technologies, then all of a sudden everyone's running for the latest, greatest little thing, even though a lot of those companies become features in someone else's product eventually.

Right. Right. Either that's just the way of the world.

Yeah. Either by copying the future or by acquiring the company, acquiring it, some of those Yep. Right.

Big fish, eat little fish and so on. So, uh, so, so from what you're telling me then, it seems like we're entering this a period of stability, and it may also be macroeconomic conditions right. At play where people are looking, you know, they don't want to take a chance on a company that may not be here a year from now or, you know, is not, is not profitable, it's not financially stable, et cetera.

Yeah. And I, I think a lot of companies are looking to optimize their observability investment, and you can save money by consolidating, No doubt, no doubt about it. So that, that's a, a big, you know, kind of realization.

What else kind of pops from the, uh, forecast this year, Peter, that you want to share? Um, so observability seems to be on the rise from our respondents, um, which is not a surprise. Um, I think, you know, you mentioned earlier ability is seen as, you know, part of the way that, that people build software today, but we specifically saw that full stack observability in the way that we define it, you know, those five elements of observability.

It actually increased 58% year over year from respondents, which is a pretty big increase. And what that tells us is that people may have had blind spots in prior years where perhaps they're not, uh, they don't have a log management solution, or maybe they're not monitoring their clients. Um, you know, so they don't have that full stack observability.

And we define seven team different observability categories in those five main categories. And the respondents told us that by the end of two years from now, so by the end of 20 26, 80 2% of organizations, and this is a number that I'm not sure if, if we can, um, it's a surprising number. I'll just say that 82% of respondents expect to deploy all 17 observability capabilities.

So they don't have it today really, But in the next few years, 82%. So that speaks to the value, I think that people understand that they should be monitoring and observing their software more, and it's just a matter of getting the right solution and rolling it out, You know, so that 82% Peter fits really well into the, you know, the whole crossing the chasm Yeah. Uh, model, right?

You get your 15% early adopters, uh, was it 30 or 35% early mainstream adopters, 30, 35% later mainstream adopters. And then there's the 15% laggards, right, who may never adopt. Right.

There it is right there, man. Right. There are your 15% laggards or the 18% in this case, and 82% represents the market.

I mean, that's a critical mass. Anything over 70 five's critical like that. Right.

So it's, that's a, that's a classic crossing the chasm sort of adoption curve that shows look. So within two years it is, it is the defacto standard that you, you do these things. And I, I think that's because, you know, it's such an obvious r o I, not only conceptually, because, you know, engineers will tell you, Hey, we're flying blind without the stuff.

We really need to see into our systems to see how they're performing and what's happening in order to address when there are problems or to predict when problems will occur so we can prevent them. But there's also the financial aspect. One of the, um, numbers that came outta the report is that we asked respondents how much money they save for every dollar they invest in observability.

And so for every dollar invested in observability, they're getting $2 in cost savings back. Really? So that's kind of like an easy sell, not only to the engineering.

Yeah, No, that's a great r o i, yeah. In a business where it's hard to show R o i right. That, that's, that's amazing.

And 41% of those respondents receive more than $1 million in annual value. So it's a, it's a significant investment that companies are making in observability and a significant return on that investment. Excellent.

Let me ask you another question about observability in general. com. They weren't known as a big open source company, but you guys, you know, promote or, or support, uh, open Telemetry, I think Prometheus as well.

Um, and yet today we we're hearing a lot of, there's a lot of people saying, well, this open source stuff isn't all it's cracked up to be in terms of security, in terms of management and vision and, and what have you, do you think that has any bearing on, on not only just New Relic, but the market in general? Well, we believe in open source and also open standards, but we also think that if you don't have a vendor behind giving you the support, the infrastructure and you know, the, the knowledge, then you're at risk. You're taking on more risk than you probably need to be, because whatever your business is, you don't also need to take on the risk of being an observability provider.

So I, I'm not sure why companies would wanna take that on instead, it's sort of, it makes more sense to me, and I think for a lot of our customers to rely on the expertise that we bring to Bear. You know, new Relic is an observability company at scale, we will ingest about three exabytes of data this year. So it's a tremendous amount of data trivial Expertise Behind those systems, right?

But, and yet we, we do embrace open source in that all of our agents are open source. So if you need or want to make, uh, a modification to them, you're welcome to. We also embrace Open Telemetry, and in fact, we're providing a path for customers to adopt Open Telemetry because customers tell us they don't want to be forced to have vendor lock in over the long term.

And Open Telemetry is a way for them to take control of their own destiny. So we want to win customers because we've got the best solution, not because we've tricked them into adopting our system, and we've got our, you know, tendrils into you. So we're adopting Open Telemetry in a big way, perhaps bigger than any other observability vendor, and we're giving customers a path to get there.

That's painless because we're incorporating Open Telemetry into our existing agents. And so we're providing a distribution of open telemetry that contains not only our capabilities, but also open telemetry. So when you're ready, just to upgrade your A agent to the latest version of New Relic's agent, and then when you're ready to move to Open Telemetry, it's literally just a configuration change.

And all of your dashboarding, all of your a p m, everything will continue to work exactly the way, uh, that, that you expect from New Relic. I love it. Peter, we're almost outta time.

Anything we left out, uh, on this, uh, forecast that you wanna let people make people aware of? Um, I think we, we nailed most of it, but I'd, I'd love to come back for a second to one of the topics that we talked about at the top of the show, and also a little bit off camera, which is, you know, some of the trends that were uncovered in the report, and one of the largest ones is generative AI and its impact on the industry. So we see, you know, at New Relic, we see the impact of Gen AI occurring in three distinct buckets.

Uh, and I'll go in order of least interesting to most interesting. So the least interesting is that in order to understand this technology, we have to be using it and be at the forefront of it's, you know, stretching our legs and, and pushing its boundaries. So we're taking a very aggressive stance to use Gen AI in our day-to-day work.

And if any listener out there or anybody watching this program hasn't tried to use it on a daily basis, I I encourage you to try it. It it'll blow your mind. Mm-hmm.

And there's so many different applications. The second bucket of value, which is the one that I'm most directly involved with, I'm super excited about, is that we're changing our user interface to provide a sidekick experience with something we call New Relic Grok. So right beside your charts, your graphs, your a p m, your curate experiences, your dashboards, you can open up a side panel and have a chat based conversation with our Gen AI agent called Grok and have a conversation about whatever you're looking at.

So if you're looking at some production service, you can ask questions about it, you know, when was the latest deployment, what's happening to the error rate? And we can give you tremendous insights based on generative ai. I love that.

I love it. That's very cool. Now, it's not the time, but I'd love, we gotta sit down.

I wanna understand how you guys are doing, doing or maybe a, a custom l l m or so, but we'll, we'll talk off camera about it. So that's number two. And then the third one is much like we're doing in our, in with New Relic Rock, there is a number of companies, in fact, the, I would argue the majority of companies out there right now are trying to figure out their gen AI solution.

And much in the same way that everybody rushed to the web, every business needed a website and then everybody rushed to mobile. Everybody needed a mobile app. Now everybody needs a Gen AI solution, and it's got a unique set of observability challenges so much in the same way New Relic helped people move to the web, helped businesses move to mobile now.

So to will New Relic help companies implement their Gen AI solution by providing the observability they need to maintain great response rate, good performance, and manage their costs, because we can automatically, a lot of people are using third party vendors to provide the answers and the insights with open AI or, or whichever vendor they're choosing. And so we can help them not only monitor the interactions, but the overall cost of using those services. I love it.

There's a lot of companies kind of growing their own stuff too, and it, we, you know, Peter, we, uh, about a month ago now, we had a hackathon down here in the offices actually, and we had about a dozen 14 people, kind of founders of the DevOps room. I had Patrick Dubois who came up with the term DevOps, right? John Willis, Damon Edwards kinda started DevOps days here in the us.

Uh, Shannon Le who started DevSecOps, Tracy Bannon from Mire, uh, mark Hinkle, what about a dozen serious DevOps ops people who are really got the AI gen AI bug. And they spent a week coding up, let's call it operationalizing ai, AI for ops, and, um, um, talk about Blow Your Mind stuff they, I saw here, by the way, we, we just, for our audience, we filmed everything and we're chopping it up into like watchable chunks right now with narration and stuff. So you'll be able to see some of the stuff we've they did, but Peter crazy, crazy.

Like, I I, it, it really blew my mind. So yeah, I, I think I guarantee you when you're on here, God willing, we're all here for New Relic 2024 Observability Forecast. This is gonna be the story.

So yeah, you know, you heard it here first actually, we have a show in January. We have a virtual event predict where we're predicting, you know, what's gonna be the big, and obviously you don't have to be a genius to predict Gen AI for next year, but love to have new relic's thoughts on 2024 observability trends at Predict, and then we can match it up 10 months later with whether it was real or not. That's good.

You know, cool, man. Sounds like a point. Hey Peter, thank you so much for coming on and spending your time with us here today on Textron tv.

For those who, as Peter mentioned, all of the data on this report is there for you to just download and, and, you know, slice and dice it to your C Fit. Or you could take the New Relic report itself, the forecast itself, and, and see it there. com.

Peter, don't be a stranger. Come in, visit us. I hope to see you soon.

Alright, thanks for having me, Alan. Take care. Alrightyy Peter Azaris, chief Strategy and Design Officer at New Relic here on Textron tv.

We're gonna take a break. We'll be right back.