현재 이 페이지는 영어로만 제공됩니다.

Observy McObsevface graphic with image of Austin Parker

In this episode, Austin Parker, Principal Developer Advocate at Lightstep, talks to our own Jonan Scheffler about OpenTelemetry, an observability framework for cloud native software that is a collection of tools, APIs, and SDKs. You use it to instrument, generate, collect, and export telemetry data (metrics, logs, and traces) for analysis in order to understand your software's performance and behavior.

Parker tells Jonan how he started/threw together Deserted Island DevOps (we were a proud sponsor) during the height of the first wave of the COVID-19 pandemic, and how overwhelmingly successful and loved it was. DIDevOps was a single-day, virtual event that was livestreamed on Twitch in April. All presentations took place in the world of Animal Crossing: New Horizons, and the event pulled 15,000 unique viewers on the day of the conference alone!

Should you find a burning need to share your thoughts or rants about the show, please spray them at devrel@newrelic.com. While you’re going to all the trouble of shipping us some bytes, please consider taking a moment to let us know what you’d like to hear on the show in the future. Despite the all-caps flaming you will receive in response, please know that we are sincerely interested in your feedback; we aim to appease. Follow us on the Twitters: @ObservyMcObserv.

Jonan Scheffler: Hello and welcome back to Observy McObservface, the observability podcast we let the internet name and got exactly what we deserve. My name is Jonan. I'm on the developer relations team here at New Relic, and I will be back every week with a new guest and the latest in observability news and trends. If you have an idea for a topic you'd like to hear me cover on this show, or perhaps a guest you would like to hear from, maybe you would like to appear as a guest yourself, please reach out. My email address is jonan@newrelic.com. You can also find me on Twitter as @thejonanshow. We are here to give the people what they want. This is the people's observability podcast. Thank you so much for joining us. Enjoy the show.

I am joined here today by Austin Parker. How are you, Austin?

Austin Parker: I'm great.

Jonan: Austin works at a little company called Lightstep that is working on this thing called OpenTelemetry—seems tangentially relevant to observability in the way that it's flipped all of observability upside down, I think.

Austin: Yeah, tangentially relevant—good.

Jonan: I don't know if upside down is the right word, but it is definitely the future of this. I worked at New Relic for a long time, about five years ago before I'm back here on my second lap. And even at the time, the engineers I worked with were constantly frustrated by the protocols and standards that we were using. And then interoperating with other people's software, open telemetry changes, all of that. I wonder if you could do a simple breakdown—a layperson’s breakdown—for our audience here if they happen not to have heard of it.

Austin: The simplest way to describe it is, OpenTelemetry is an observability framework that you can use to integrate it into your application code, service code, but also into your infrastructure. So you can integrate this into Kubernetes. You can integrate this into Lambda functions or serverless or whatever else. And you can use it to generate distributed traces, metrics. In the future, you'll be able to use it to grab your logs as well, and then send them to any kind of analysis system you want. So one of the great is, it standardizes a lot of the conventions and a lot of the data formats and on the headers for propagating context. And it says, “Hey, here's sort of what we can all use to talk to each other,” HTTP server and Ruby is going to have the same sort of attributes as an HTTP server and go or C# or Java so that I can have these nice semantic conventions for what things look like, and by analysis platform.

And there's a lot of tooling built up around it to make it easy to collect all this different telemetry from applications, from infrastructure, from wherever, transform it into other formats and then send it off somewhere. So for example, you can use the OpenTelemetry collector to take your legacy Zipkin traces, and then your Prometheus metrics and turn them into a different format. You can turn them into the new open telemetry protocol and there's several platforms that accept that. You can also just turn them into data. You can turn them into storage, into JSON Blob effectively, sit on Blob storage like GCs or S3 or whatever, and write queries across them. We saw this week that Grafana is doing something called Tempo, which is with our new trace data store. And it basically just takes traces and puts them into Blob storage bucket and does some other fancy stuff. And that's built on OpenTelemetry, OpenTelemetry’s data format and OpenTelemetry idioms and protocols.

Jonan: The project's called Tempo. So it's a data store specifically for traces.

Austin: Yeah.

Jonan: Interesting.

Austin: Yeah. They just announced that their observability can—and I know New Relic is a huge contributor to OpenTelemetry as well and is planning on supporting it. I think I read summer of 2021 would be full support.

Austin: I read a blog.

Jonan: Yes, here's a blog.

Austin: There was a public blog about it that I remember reading. So…

Jonan: Yes, I know that we are in OpenTelemetry already with many of our agents when we open source everything, a lot of it was already over to using, or heading towards OpenTelemetry. I just don't know what the timeline is really.

Austin: Yeah. All I know is, I remember reading, at least there was a projection on the blog, but yeah, definitely besides New Relic, besides Lightstep, obviously you see Grafana getting into OpenTelemetry—Datadog has a lot people contributing to the project, Honeycomb, Splunk…

Jonan: Is Prometheus working with it yet?

Austin: Prometheus works with it, yeah. And that's the great thing about the way OpenTelemetry works, is the idea is it's very agnostic to where the data goes. So you have these collectors that can take data from any source that has a receiver and then transform it into anything that ExFolders. So you can take metrics from five different types of things, and then you can turn them all into Prometheus metrics if you want to and send them to Prometheus.

Jonan: Right.

Austin: And that kind of Swiss army knife is, I think, really useful because a lot of times what you see out there in the real world is, you have different teams using different things, depending on when something was written or how well you were doing monitoring and observability at a certain point in your product or lifespan. You might have four or five different things and you've got some dashboards somewhere that tries to consolidate all this stuff, but it's a lot of clicking between things. A lot of not knowing where the right information is.

Jonan: And trying to sew it all back together yourself.

Austin: Right. And with OpenTelemetry, that process at least should become a lot easier and give you a good place to build on for the future. Right?

Jonan: Yeah. I am excited most about that interoperability piece. I think the mentality we saw in software 20 years ago, even 10 years ago, around building up these walled gardens and making sure that your moat existed by making your product difficult to integrate. I think that that way of thinking is going away, and I'm really glad to see it die. Having spent a lot of time this week, trying to get an Ext. for an extended for journaling?

Austin: Yeah.

Jonan: Possibly working on my Mac. Yeah, I'm over it. I'm good on that.

Austin: I also think it's a realization, I would almost say, from a lot of the different vendors and people that have been big players in this. It used to be, everyone had their own agent and those agents weren't necessarily compatible with each other. And what have you seen with OpenTelemetry? You've seen a lot of those same companies saying like, “Hey, we're going to open source this, we're going to donate this OpenTelemetry,” or “We're going to have engineers work on making just a single agent.” There's a really, really good Java agent for open Telemetry, for example, that does what you'd expect—you plug it into your process and it will go through and instrument your code, and it's great.

But I'm hoping that that sort of open source, automatic instrumentation where you pick and choose, you don't have to feel like you’re in a walled garden. I can just use this kind of standard, and then I can make my choice based on what's the best platform. And start thinking, “OK, how do these things interoperate? How can I get data from one place to another and make that data more useful outside of just developers.” Because I like to say, “Observability is for everybody.” It's not just an engineering thing.

Jonan: Yeah. And it has long been exactly that. I think a lot of these companies that exist in the space today rose to prominence because they were able to give detailed insight. There was, to some degree, these kinds of dashboards you could put together for your executives or other people within the company and provide that data. But it has always been based on the agents. You're right. And now I wonder what becomes of the world. It reminds me of the multi-cloud kind of thing. Not in that multi-cloud is a good idea, but that people are really motivated to achieve it because it commoditizes the platform.

So you often have situations where Google comes to an engineering leader and says, “Hey, we'll give you $2 million of free Google Cloud this year if you sign a two-year contract. And then if you have a system that is portable enough, you get to take advantage. It's like rolling your credit card debt to the zero-interest rate all the time. It quickly lands in a place where the platform becomes less relevant. And then what value can you add as a company? I think that in the case of these kind of observability companies, the value then comes in what you can tell people from the data. I don't necessarily get value from knowing it's broken. I get value from knowing why, and quickly being able to diagnose that issue in a microservices architecture, this method in this Rails app, because of this gem conflict is what is breaking my infrastructure. That's what I want to know.

Austin: And I think we talk about that a lot at Lightstep. We talk about this idea of what's changed, and it's on two levels. If you look at what is it to be a software developer in 2020 versus 2015, a lot has changed and how we build software just in a really simple level. We have better tools. If you think about DevOps, that's kind of this thing that has momentum behind it. We're very much moving towards a general agreement that you don't want to make your DevOps processes, your CI, your CD, all of that dependent on a specific platform or a specific provider or whatever.

So there's still a lot of different CI/CD tools. There's a bunch of companies. And then every single cloud provider has their own thing—but more and more, what's the most popular way to build—the most popular sort of build tools? It's probably still Make. Everyone's just kind of using Make, or maybe they're using Bazel, but you still see stuff like Terraform. You see these abstractions over platforms over, I mean, abstractions over abstractions really, that allow your entire DevOps, all your DevOps stuff to really just be separate from any one thing. And I think, or what I would least like to think, is going to happen is we're going to see that same sort of movement with observability and monitoring—where it's less about, “Hey, we're on New Relic and everything's on New Relic,” or, “Hey, we're on Lightstep, everything's on Lightstep,” or, “Hey, we're on Honeycomb, everything's on Honeycomb.” All these different companies. And it becomes more like, “OK, we're using OpenTelemetry and maybe some teams are putting stuff into Lightstep, because they have a feature that we couldn't do without, they have a really cool feature.

Some people are putting stuff into New Relic and go over the space of the Honeycomb. And then all of these platforms can output the data, output the result of that in a common format. And we can build our nice dashboards and whatever we want to build them in. Different people are going to have different needs. And the technology needs to respect that.

Jonan: Yeah. I think that's a thing that software companies forget a little too often for my taste, is that your product, no matter how good it is, is not going to suit every use case all of the time. There seems to be this assumption in tech sometimes that you can just win the entirety of a market and in a space as crowded as ours certainly, or even in other arenas as well. It's just companies sometimes neglect to understand to their detriment how important interoperability is for their success.

Austin: Yeah. I would also say it's not maybe a little unfair to say that companies forget. There are powerful financial incentives attached to winning. And a lot of those incentives come from the investor side, the business side of it. The investor side of it, you're not going to go down to Sand Hill Road or whatever and get a million or a hundred million bucks unless your story is, “Oh, we're going to crush everyone else.”

Jonan: Right. And I get that. I think that there are ways to achieve that while embracing open source is what I'm learning.

Austin: Yeah. You know, one of the things that we've started out at Lightstep as a company was, “Hey, the sort of the top of the funnel, right at the very top of this, the developers that are in the Ops people and whoever else that are trying to get telemetry data out of their software, out of their infrastructure and be able to do something with it, all of that needs to be completely open source. We've never had a proprietary agent, it's all been built on OpenTracing and now on OpenTelemetry. And that, I think, has been super important, because there are people that have tried us out and then went with something else—because they were able to do that, because all the instrumentation was built on OpenTracing and it's very easy to switch. But I think it works in both directions. Not everyone is going to have the same needs as everyone else at the same time. So if you have that open source instrumentation at the very top layer, then you can make the right decision for your business and for your sort of situation and not feel, not feel bad about it. Not feel acrimonious and not get pushed into something you don't want to, or that it doesn't work for you.

Jonan: The part where you end up trapped in some platform and then your impetus to moving is it is expensive or time-consuming, and your life is progressively harder because you're stuck on these older platforms—it's kind of like the transition to cloud. Like there are many large enterprise companies who have, in some cases, legitimate reasons, but less and less not to be in the cloud, they're still running on-prem servers. And that kind of trap that you develop for yourself, I think is unlikely to come to exist because of projects like OpenTelemetry, which I'm very excited about. Speaking of things that are exciting, I want you to tell everyone about the best conference I saw all year, Deserted Island DevOps. It was amazing.

Austin: Thank you. So Deserted Island DevOps was about the beginning of March or something—the last time I got on a plane—so it was very much in that first few weeks of shelter-in-place, the world was changing at the end, middle end of March. And I think we all had a lot of time on our hands, just as a world, as a planet, and after a couple of weeks and everything was shutting down and, we were looking at KubeCon being canceled. Basically, every event I've been booked for the remainder of the year was saying, ''Hey probably not going to happen. We'll let you know.'' I made a joke on Twitter, that was like, because I had also just gotten Animal Crossing: New Horizons, which the cultural game of the year, I'm not sure, like everyone, if you don't know what Animal Crossing is, you don't use Twitter at least.

Jonan: Yeah, you're probably not on Twitter.

Austin: So I had Animal Crossing and one of the cool things in Animal Crossing is you can design your own patterns. And so I went in and I made a fake little booth in there with some, put a laptop on a stand and put some of the pixelated glasses on the thing. And I'm like, take a screenshot and like, ''Oh, I re-created a conference booth, recreated a trade show booth inside animal crossing. And I put this out on Twitter and a couple of people came and were like, ''Oh, that's funny.'' And someone was like, ''Oh, I would go to a conference if it was an Animal Crossing.'' And so I was like, ''Huh, well, how would you do that?'' So one thing kind of led to another and got the idea of like, “OK, well, you know, I'll make a webpage for this because why not? We do an Animal Crossing tech conference?” I made a webpage, I put it up on April 1st, a couple of days after this tweet and put a link to register for this mailing list. And I'm like, “I'll put it out there in the world. And if nobody comes in and signs up, then it'll be like, “Ah, ha ha, yeah, funny joke.” Well, it got really popular. And I had, I think, like a hundred people sign up just within a few hours. I'm like, “Oh, so there's actually some interest in this.” And that was kind of how it started.

It was one-day. We streamed for about six, seven hours. It's kind of what it sounds like. It was a dev, you know, if you've ever been to a DevOpsDays or a tech conference in general, it's probably more equivalent to like a DevOpsDays or a meetup. And we had people give presentations on everything from, Ian Coldwater did a great keynote on security and devs and security and how to interact between your security team and your developers. We have people talking about mental health and burnout, taking care of yourself during a pandemic. We had some great talks on, interestingly enough, on what tools you should use, when's the right time to go with the new thing versus the old thing. All these videos are up on YouTube if you want to go watch the actual talks, they were all fantastic. And the way it actually worked is people would view this on Twitch.tv. So we had a Twitch stream.

What people saw on Twitch was, I was sitting on the Island and I had the camera and I was kind of my perspective, but people would go and they would go up to the little podium and they would share their slides of resume. And I, on my end, overlaid the slides with the feed of Animal Crossing and then sent all that out to Twitch. And we had a discord where people could like chat and ask questions. I think we had 15,000 unique viewers on the day of, it was like the third or fourth most popular thing in the Animal Crossing category on Twitch for a little while.

Jonan: For comparison sake, for people who are unaware, 15,000 simultaneous viewers on Twitch is huge. I mean, you had to be one of the top streams on Twitch at the time.

Austin: So it was like 8,000. It was 15,000 unique, but we had over the course of the day, but, actually, I think our max was 9,000, which is still pretty high.

Jonan: Crazy. Yeah.

Austin: Especially for a tech conference in Animal Crossing.

Jonan: Amazing.

Austin: The entire thing went from idea to happening in 30 days.

Jonan: Wow.

Austin: It was very much a throw stuff at the wall and see what sticks approach and it worked, but it worked out very well. I think people still talk about it.

Jonan: It was by far my favorite conference that I've seen. I've been really disappointed in how the events industry has responded to the pandemic. This new reality, where we are now all trapped in the richest multimedia experience that has ever existed. And we are using that to build three-dimensional event halls with smaller slideshow real estate than I experienced in an actual event hall. And I was really impressed with the interactivity that you could offer.

Austin: There really are two or three big takeaways. One is, you like having an audience when you're giving a talk, because the audience not only gives you feedback, active feedback on what you're saying. Are my points landing? Do people understand what's going on? But also, you're talking to someone. There's a huge difference between if I was sitting here and we weren't on a Zoom and I couldn't see your face, I couldn't get cues.

Jonan: Yeah.

Austin: That would be a very different experience to just talking to a camera and there's my slides and nothing else. And so many of these virtual conferences that you see people put on, it's all, “Oh, you got to prerecord your talk, dah, dah, dah, dah.” And in that case, you're hitting record and you're talking to your camera and you're giving a lecture, as opposed to having a conversation. The kind of talks that I like to give and the kind of talks I like to go to are conversations. So one of the things that was really useful about this is that we did all the talks live and we had all the speakers in a Zoom together at the same time. So when you were speaking, you saw all the other speakers in that Zoom. You could see their faces and see them react to what you're saying, but you could also see we have the Twitch chat going one, people would be clapping. They would be emoting. They say, “Hey, good point.” “Hey, Oh, that got me thinking about blah, blah, blah, blah, blah.”

And that is the most important part—that immediate feedback of like, “Yes, people like that.” That's what's missing so much in giving a virtual talk, even a live virtual talk to a normal conference audience, you got to have that sort of connection, that way for people to really quickly say, “Yes, I liked this and I don't see anyone working on that.” It doesn't seem to be what any of these platforms are really optimizing for, which is just disappointing to me.

Jonan: I, as a presenter, play off of the audience's reaction so much. I miss it desperately being able to see when I've lost people's attention and when I need to switch it up or I move off on some tangent and I pull people back at the right moment before I lose them all again—that ebb and flow of the conversation format. I have not seen it re-created well, but that Zoom feature where you had a virtual greenroom going effectively, I have heard about recently on this podcast from Matt Stratton, I think, if I remember correctly.

Austin: Yeah, they did that for DevOps Chicago, I believe.

Jonan: Yeah. It's a good approach. And I think it's just the beginning of where people can take these things. There have got to be better ways to get engagement than clicking into a virtual exhibit hall and dropping by the thing.

Austin: I am dev rel person, but I am not unsympathetic to the business side of this. Why does anyone pay enormous sums as a sponsor, even an in-person event or a virtual one? It's for business value. It's not just, “We want people to see our logo.” You know, they have salespeople that want to get leads, and they want to talk to attendees and dah, dah, dah, dah. This is my experience, and other people might disagree or have a different one, but I've attended quite a few virtual events this year, and I have never seen a single person go into a virtual booth and strike up a conversation with a salesperson.

Jonan: Yeah. Some of them have a point system, so you can get swag mailed to you if you go to enough booths or something. And so you've got people dropping in and saying, “Thank you for coming.” And then leaving immediately.

Austin: Right. It's even more transactional than a normal sort of booth thing is. I think a lot of that is actually because of the way advertising works. If you're on a conference floor, there's actually two parts. One is the way advertising works. If you're at a conference and you're walking around with your coffee and muffin, you might see a looping demo or you'll see something—a booth, a piece of art. And then you're like, “Oh, huh, that's interesting.” And you go over and then chat with somebody. Or maybe you really do want a pair of socks and you hear the pitch. And you're like, “Oh, that's interesting. That reminds me of something.” And you get into a conversation and it's actually a valuable interaction. And that just doesn't happen with a virtual booth. There's nothing that actually grabs your attention. There's no stick, effectively.

Jonan: Yeah.

Austin: At a real in-person thing, you're out of your normal day-to-day, and you're in a weird liminal space of a conference. And so you will do things that you would probably never do unless you're at Costco. And at Costco, you see a sample thing and you'll go over and you'll get the hotdogs or whatever, and be like, “Oh, I'll buy a pack of these hotdogs.” But if you're online, how many times has anyone in the history of time ever seen the people also buy these things and click, “Yeah, I'm going to do that.”

Jonan: Yeah.

Austin: You know, it's such a different mode of interaction, and we don't have strategies for it and we don't have ways to deal with it.

Austin: So my thinking is, as an industry, companies would be more OK with sponsoring if the ask was less. Less money, but instead of doing booths—instead of saying, “Oh, you'll get leads or whatever”—just run commercials, just do tight 30 seconds. And if you go, “Here's a webpage, maybe if you go to our webpage and fill out a survey, we'll get you a thing.” But stop sending sales reps to have another Zoom, another monitor for eight hours a day. It's not doing anything for anyone.

Jonan: Yeah. I don't think it's adding a whole heck of a lot of value, but I do think that there is room there. I think there's also the piece of a live conference that, you talk about, people getting out of their comfort zone a little bit. There's just the novelty of the experience. And when I have an online event, I don't have that sense of urgency. I'm there in Cincinnati for this conference or wherever I am, and I'm not going to be back there for a couple of years. You know, even if I am in Devereaux, it's going to be a while before I come back on this lap. So I want to be there and experience that. And I'm in the hall with all of my friends and then I'm online and all the videos are going to be recorded or you're going to gate them behind a paywall, that's real messed up. I get frustrated with that, but I know it's there.

Austin: Yeah. And for someone that's not in dev, right, just someone that's trying to go and learn and do whatever, and they're working from home. Gosh, it's probably even worse because other stuff happens.

Jonan: Yeah.

Austin: Are you really going to sit there for eight hours and just watch a bunch of videos on YouTube? Do you have other work to do? I don't think people respect your time. I think there's probably something people haven't quite figured out—like, how is it even possible to disconnect from your job and from like Slack and from Jira and stuff like that when you're in the same place. Maybe you do a little mini-vacation and you take your laptop to the living room and you watch the videos there. I don't know. There's more questions than answers.

Jonan: Yeah. I think we have a long way to go in the online events, but hopefully we'll just have the real ones back again. I miss travel so desperately and seeing other humans and hugging them, just hugging other people. I have my family now, but I miss hugging my conference friends.

Austin: Yeah. I will say this. I think when they come back, we will see hybrid events, certainly by the biggest of the big. Your re:Invents, your Builds, your Salesforces [Dreamforce], your significantly large companies that are doing end-user events or developer events or whatever. I think they'll probably be doing like fully hybrid ones, because from an accessibility and an inclusiveness perspective, that is super valuable.

Jonan: Yup. 1,000% agree. And I think that the hybrid event makes perfect sense, in part, because what you're doing is streaming live content—live video is just more engaging.

Austin: It is. And I think if you see a way to make those live video participants on the other end of it feel like they are there too, and make sure the speakers at the hybrid events can be aware of them and understand like, “Yes, I have these two, I have the audience that's in front of me. And I also have the live audience, the live remote audience.” And they're both important. That's a solvable problem. And I think that's a problem that will honestly make for better events. Now I would also say that I see some of these—I don't know—I'm planning on doing another Deserted Island DevOps next year.

Jonan: We're planning on sponsoring another Deserted Island DevOps next year, I would like to know when your CFP opens as well, please.

Austin: I'm not going to say any hard dates, but look for news about that in the new year.

Jonan: Excellent.

Austin: But that said, it'll be a little different than last time. But I do I see a world where I keep doing that even after the pandemic is over? Probably.

Jonan: Yeah.

Austin: There's a lot of intangible things about a virtual event that tries to be more than a virtual event. And the thing that this has disappointed me the most about the year is that, like you said, we have a golden opportunity for the events industry to step up and be like, “Hey, we're going to make things that are really, really special and different and are not leveraging the platforms, leveraging the technology we have." And it just isn't happening.

Jonan: People have been busy, it's hard right now. I've had a couple of events waiting in the wings for months now. I think that you get out of bed today and you change into real clothes, and that's a win. As you know, especially for people who aren't used to working from home, I've been doing it for a lot of years, and there are tricks you pick up. You don't put Slack on your phone, you have a separate laptop with none of your work stuff on it. It doesn't really even apply in this period of time. This has been by far the hardest period of working from home that I've ever had, just maintaining that motivation and the separation of work/life balance because given the opportunity with everything going on in the world, it's real easy to just go face down into your work and disappear in it all. So Deserted Island DevOps is coming up and people can find you online and watch the videos there. What other sort of places on the internet could people come and find you and check out what you're up to?

Austin: Oh gosh. So I tweet @austinlparker. I tweet a lot. So that is the authoritative source for what is going on. Other stuff coming up. If you're going to KubeCon and you're interested in observability and OpenTelemetry, specifically, registration is still open for OpenTelemetry Community Day. That's going to be on November 17th, it's a day zero event, or it's a co-located event with KubeCon virtual. That actually should be pretty interesting, because we're not doing a traditional virtual conference, it's an unconference. So the whole thing is kind of like birds-of-a-feather sessions that you, as the attendees, figure out what you want to talk about and go off and have your own discussions, talk to maintainers. And we're trying something a little bit different that is less like a Zoom, less like an hour Zoom and more like hanging out with people and talking about cool stuff that's coming up. And talking about how you can get into OpenTelemetry, how you can use observability, things that have worked for you, things that haven't worked for you. So keep an eye out for that.

Jonan: The open telemetry community unconference, what is it called?

Austin: Open Telemetry Community Day.

Jonan: Open Telemetry Community Day. Excellent. I will do my best to get in there, that sounds lovely.

Austin: Yeah, it's a $20 registration fee on top of your KubeCon registration.

Jonan: So your blog.

Austin: Yeah. I blog very infrequently at aparker.io.

Jonan: Oh, I forgot a step. I have a tradition here on the show of getting people to make a prediction, so we can have them back a year from now and be like, “I cannot believe that you said Microsoft was going to win the cloud.”

Austin: Oh gosh. That's a tough one. There's so many things that you could make predictions about that I don't that feel like…

Jonan: I have one.

Austin: If you speak them into the air, you'll change things.

Jonan: Yeah. We're not going to predict anything that might change in the next week or so, but a year from now, you'll be on Deserted Island DevOps 3. I think within a year, you'll have your third version.

Austin: OK. I predict that by this time next year, we will have had Deserted Island DevOps 2021. And we'll see where we go from there.

Jonan: I really like your Twitter profile—"the Jason Mamoa of DevOps.” Please don't ever change that.

Austin: Thank you. I have a regular podcast with a couple of other people that is definitely not a work or a podcast, but one of my co-hosts made that joke and it stuck.

Jonan: I love it. Well, thank you so much for joining us. Do you have any parting words of wisdom? Maybe for people who are a little earlier on in their DevOps careers, tell them all to study up on OpenTelemetry? [laughs]

Austin: I'll tell you this—gosh, five years ago would have been 2015, what was I doing? Maybe not five years ago, 10 years ago, I was probably answering phones at a call center or working fixing cell phones. So you gotta be lucky and you gotta know what you're worth. I mean, I guess the biggest thing is, don't get knocked down without getting back up. Right?

Jonan: Yeah.

Austin: And be aware of what's going on around you. They said that running an Animal Crossing DevOps conference was really silly, and they were right. But it's also turned out to be a really good idea. So don't listen to what they say.

Jonan: To you.

Austin: Yeah. Don't listen to what I say either.

Jonan: [laughs] That's really good advice.

Austin: Yeah. Then use that information in the best way possible. This is the first time I've ever been on a podcast that someone has asked life advice from me, which I don't know if that says more about the type of podcasts I go on or just people's assessment of me as a person to give life advice.

Jonan: No, I think it's just, I maybe am less interested in having a traditional technical, deep dive of a podcast than I am about having human beings talking about human beings.

Austin: That's a good point.

Jonan: Yeah.

Austin: I've been a union organizer before, I've worked as a short-order cook. I've worked at a newspaper. I've done web development. I've done a lot of different things. And I know that a lot of the opportunities I've had in life have been like, “Oh, I'm, you know, a normal looking white guy.” And so that certainly helps. And not everyone has that option, but I think a lot of it also comes down to, you have to be able to be thinking about what's next and be there when you can kind of make those jumps. You can do that wild, unexpected thing and maybe it'll work, and maybe it won't work, and you got to get lucky and you have to be good. But if you're not in the right mindset or you're not able to make those decisions, then that can be tough. So you have to be willing to take risks, and I get that not everyone can. I don't have a great answer for you other than smash the state, but yes.

Jonan: [chuckles] I think that is a good closing argument right now that I will take.

Austin: All right.

Jonan: Again, thank you so much for your time. That was lovely. And I look forward to having you back in a year, just about the time that you announce Deserted Island DevOps 3.

Austin: Three. All right. I will look forward to it. Thanks so much for having me on.

Jonan: Thank you so much for joining us for another episode of Observy McObservface. This podcast is available on Spotify and iTunes and wherever fine podcasts are sold. Please remember to subscribe so you don't miss an episode. If you have an idea for a topic or a guest you would like to hear on the show, please reach out to me. My email address is jonan@newrelic.com. You can also find me on Twitter as @thejonanshow. The show notes for today's episode along with many other lovely nerdy things are available on developer.newrelic.com. Stop by and check it out. Thank you so much. Have a great day.