PurePerformance - Whats Hot in Cloud and AI-Native and what we learned from the AWS Outage

Episode Date: October 27, 2025

The AWS US-East problems on Oct 27th was a good reminder how depending we are on globally shared services. Built-in Resiliency is not guaranteed if systems have a hard dependency on a single region of a single vendor. Many of us have experienced systems being impacted that we use on a daily basis - some critical - some not so critical as Andi will tell you when he found out that is beloved Leberkas Pepi App didnt work!Besides this outage we discuss lessons learned from Cloud Native Days Austria, Observability and Platform Engineering Meetups in Gdansk and Tallinn as well as giving an outline to the upcoming Cloud and AI-Native US Tour from Henrik Rexed and Andi GrabnerAll the links we discussed are hereLeberkas Pepi: https://www.leberkaspepi.at/Cloud Native Austria: https://www.linkedin.com/company/cndaustria/Observability Meetup: https://www.meetup.com/observability-tech-community-meetup-group/US Tour from Henrik and Andi: https://events.dynatrace.com/noram-all-de-engineering-efficiency-tour-2025-28225/

It's time for Pure Performance. Get your stopwatch is ready. It's time for Pure Performance with Andy Grabner and Brian Wilson. Hello, everybody, and welcome to another episode of Pure Performance. My name is Brian. Wilson and as always I have with me my co-host and special guest today who's making head gestures he's not mimicking my talking but he's turning his head and and being fun Andy Grabner ladies and gentlemen the guy you probably never heard much before but here

he is I'd like to introduce you and Andy why don't you tell the audience a little bit about yourself today well thank you Brian so much for having me and yeah for those people that don't know me most likely because they've never listened to this podcast before. My name is Andy Grabner, and I think we are probably, we're hoping for different guests, but this is just our faulting that we just interview each other. Yes. We do have a bunch of guests lined up, but we were going to have a gap if we don't do something like this. Exactly. But Andy's always, you know, a world traveler talks to a lot of people. So for the listeners who don't know, Andy's always getting our guests, he's our, well, he's the brain trust behind

this podcast, right? Not only is he really smart on all the topic. but he also is very well connected and meets tons of people at all the conferences he goes to so he gets to bring in a lot of people to the show that we you know we love talking to but also you know one thing not to overlook andy is the fact that you are going to all this conferences you do talk to all these people so you get a bunch of information and knowledge just from those and one of the reasons why I like when I just talk to you is that we get to just talk about what you've learned from that but before we go into what you You learned, coincidentally, something really big happened today.

Really? Something scary happened, even like perfect for a Halloween episode. Because if you look at the picture at our thumbnail, we tried to be scary. And what happened this morning. So today is October. As scary as we are funny. Yeah. October 20th, 11 days before Halloween, the world stood still, or at least,

AWS-U.S. U.S. East stood still for a couple of hours. Brian, I hope you were still sleeping because it wasn't in the early morning hours. I was hoping it would still be out. When I saw some stuff on my blue sky feed, I was hoping to see that it was still out or something because then it would be like, oh, I guess I can't work today, but it didn't get so lucky. But again, I guess I get to talk to you because we don't know what's on AWS. Yeah, true. And I think the reason why I wanted to bring it up, obviously, this is no finger pointing.

It's really, if these situations happen, whoever is responsible for any type of cloud and shared service, I'm sure they didn't have a lovely night, the folks that had to do it. Well, I don't know yet, but what I want to say is I think a couple of things. A, it shows us how dependent we are obviously on shared infrastructure and how important it is to think about resiliency from multiple angles. I think this is also where people think about multi-regional. This is maybe where people also think about multi-cloud. what if one of your clouds, that normally never goes down. But if it goes down, then what is your contingency plan?

What's your plan B? And also one of the reasons why I wanted to bring it up last week, I was in Estonia, in Tallinn, and I was presenting at a meetup. And I talked a lot about platform engineering, but the discussion obviously came out about observability. And I told, and I talked a little bit about diner trace and how we do observability and that we are providing dinotrace as a SaaS service and then somebody asked me in the audience, so how do you monitor dinotrace?

Well, we use diner trace to also monitor diner trace. And then he said, how do you monitor this diner trace? Because if diner trace goes down, how do you do this, right? And it was a really good fair question. Now, obviously, we have diner trace deployed in different regions, also across different cloud vendors. So that means the idea is, and this is also true for our customers, right? You can obviously monitor and set up different tenants in different regions.

to make sure that you are not flying blind. Nevertheless, I also want to be honest, and I think many of our of the SaaS software vendors today were impacted because U.S. East runs a lot of critical services. I know we are using ticketing, like Asana internally in my team for ticketing that was done. We are using Restream as a recording facility tool, a recording tool for videos. That was also impacted. So it had a big impact.

And I think many, also Docker had issues and everything. So this is, it's just an interesting reminder that these shared services are critical infrastructure. But we still need to make sure that we have a plan on how to deal with it when these things don't work as expected. Right, right. And not to go into the geopolitical side of it, like we're not going to dive deep, but it also just points out how reliable we are on this and obviously there's a lot

going on geopolitically you know war's breaking out everywhere and everything and like you know

you think about like all right if someone maliciously takes down one or more of these things like everything

comes to a standstill or everything I remember you know we had a a blackout I don't know after Hurricane Sandy

years and years ago in on the East Coast we didn't have power it's like you can't even really pump gas without power unless you bring in a backup generator,

but then you need gas for that right. So everything is relying on what you'd call these utilities, right? Yeah. You know, like power, water, internet. The other thing, too, is not to get crazy, but what if one of these providers decides to start holding their data center's hostage? Like, we're going to turn it off unless you, yeah. You know, we get into James Bond stuff, but we don't need to go down that side.

But again, it's like if you're not divested and spread across multiple components, like the idea you mentioned, you know, how do you monitor Dina Trace? Okay, well, we use Dina Trace to monitor Dina Trace. And then you can keep asking, like, on and on. Well, how do you monitor the monitor that's monitoring Dina Trace? And obviously, there's a set limit to that.

But if all your monitors are in either one zone or one cloud provider, you know, that defeats it. So, you know, I think it is a really great reminder that it has to be spread out, you know, depending on how critical your services are,

it is a huge risk. You know, maybe the video recording service going down, you know, for a few hours, that's not going to impact anybody, but for definitely for critical internet infrastructure components, it's huge. And even for non-critical, so a fun story on the whole thing, I went to Vienna today because I had meetings in our office in Vienna. And on the way there, I arrived just shortly before lunch, and I walked down to my favorite fast food place, Austrian fast food. It's called Levercast Beppe, so for everybody that is listening, that knows Levercase.

And so I got my Levercast email, and then I'm also on their loyalty program. So that means I have an app where you can then scan the receipt, and then you get points. And it didn't work. It just said, you know, system is currently down. Now, I'm just making assumptions, but it could very well be that the backend system, the powers the loyalty program of that local fast food restaurant, you know, runs their backend services on AWS as well. Who knows?

Right, but at least the payment service was not down. The payment was not down, and I still have to print that copy of my receipt here. Because you want your points. Because I want my points, and I want to scan it later on once the service is up and running. What do you get when you have enough points? You can buy, you can get discount on food. What's your main? Just in case anybody listening happens to run into you, and there's one of these stores nearby,

and they want them to get you a treat. What is your go-to? It's a case, lever-case, similar. So it's basically a lever-case with cheese in there. What's-le-case? Yeah, what's labor-case? I'll send a link over for everybody that is watching. Listening to this, I will definitely send a link to Leva Caspepepe.

That's the local, it's an institution here in Lince and also in other parts of the country. Yeah, but enough about labor case now, even though it just makes me hungry again. But again, this is like, you know, if, you know, hopefully we never see it. But if Azure and AWS and GCP have outages simultaneously, obviously, you know, you can only prepare for so much, right? There is a limitation. There's a cost, whatever benefit. But again, when you have those critical systems, right, you would want to, you would almost think. Maybe some of these places should also have, like, an on-prem backup, right?

I'm thinking, like, hospitals or, you know, fire and police systems, right? You don't want to just have it in the cloud, because if there is a, you know, whatever, you know what I'm saying. It is kind of scary when this happens. I haven't seen an outage. Well, I didn't see that outage, but I haven't heard about an outage like this in a long time, right? I think, you know, many years ago there was one with Cloudflare. Is Cloudflare the one with the, yeah, yeah. yeah so but um to talk about how we how do we start about this and obviously this was an event

today but you said i've been fortunate enough to travel um i was very fortunate enough to be one of the hosts of cloud native days austria that was two weeks ago in vienna and first of all thanks again for everyone in the organizer team as many people behind the scenes that are doing a lot work with a lot of volunteers that helped it was a two-day conference in a cinema in Vienna with two auditoriums a lot of helping hands and also a lot of great

speakers and I got to say for me we had this is kind of like a little bit pitch to the next podcast episode we had Laura Tako from GetTX as a keynote speaker and she talked about developer experience

for me the most interesting takeaway she had a slide and I'm pretty sure I will definitely ask her in the podcast that we record later she talked about the impact of AI on developer experience but then she brought a slide that showed what are the top time consuming tasks

that we have as engineers and Brian take a guess what do you think the number one time most time consuming task is of an engineer debugging even non-technical what do you think where do we spend most of a commuting to an office

almost there I give you a third chance now what else do we do while we're working that it does not related to writing or debugging code or testing code you've stumped me thinking

no that doesn't make sense what do you see most and you look at your calendar oh meetings yes meetings yeah meetings so but if you're not having meetings it doesn't look like you're productive

so it's a corporate must But I mean, her point that she wanted to make, there's many different things that make us unproductive or less productive and meetings seem to be the number one-time consumer. And so her question to the audience was, if you're thinking about wipe coding and using AI and the IDE, of course, that will give a productivity boost or enhancement. But are you really focusing on those tasks

that are currently consuming the most of your time? And you may not thought about where most, We may not even know where most of your time is spent. So I thought this was a really interesting takeaway from this meeting, and that's why I'm really happy that we have her. She will talk about the DXQA4 metrics. I think we had Tucson from the Czech Republic a couple of months ago, also on that topic because they have been trying to adopt that DXQA4 framework,

and now we have her. She's the CTO from GetDX, and she has a lot of good insights into how organizations measure and improve developer experience and develop efficiency. Yeah, I think it's also too, obviously meetings take up a lot of time, but every time I've either dabbled a little bit in coding or even if I'm mixing music and stuff, having an interruption is just like a huge productivity killer.

So I'll bring this up with Laura, obviously, but the idea of like, okay, if there are going to be meetings, they should probably be all stacked together at the beginning or end of the day. Or even right then it'll just put them all. together. This way, like, once you're a developer and you're working, you could just keep chugging along because your brain's going to be on that different wavelength. But very interesting. I look forward to having that conversation. We haven't had a DX podcast in quite a long time at this point. Yeah, I remember we had those, we had two son on because I met him in Prague early this year and then we talked a little bit about it. But you're right. It's been, it's been a while. Yeah. And then the other two things that I wanted to quickly highlight on that conference, the other two keynote speakers. One, It was actually the, it was David Pavlik. I met him also in Prague.

And I was then invited to go back to him inside to his company and talk about observability. But he had a keynote in Vienna that I saw that he delivered at an engineering leadership conference in Prague earlier this year. He talked about his life and his life opportunities that he had because he started his early career at Microsoft and then he basically raised his hand when the opportunity came up to say, I would like to go to the U.S. And then he kind of explained how we went from Microsoft to Amazon to Netflix to SpaceX.

And then like he showed all of his kind of paths in his career and always what triggered that opportunity and why he took it. But in the beginning and this was kind of for me the key lesson to take away is it started with raising his hand. Because the message was right, if you don't raise your hand that you want to do something, that you have a certain skill. or maybe people around you don't recognize it and put you in that position to take this opportunity. You need to be the one that raises the hand and make people aware, hey, I think I should do this.

I want to do this. I want to step up. Yeah, that's a very good point, Andy. You know, you and I have gotten, you know, pretty far in our careers. And I think it's because two things. Number one is we raised our hand and looked to do things that, you know,

looked fun and interesting, right? I mean, it's not like we were raising our hand, like, hey, the bathroom needs to be clean. Yeah, I'll do it, right? Now, this is something fun and interesting, and I want to learn about it. So you do let your intent be known so that people will be thinking about you. But the flip side of what you've got to do as well, and this is a lot harder for people like me and probably a lot of people who are developers, right, is to have some level of self-promotion, to make sure people are aware of the things that you're doing with, you know, You don't want to be bragging about it and saying, look at all those awesome stuff I've done.

But you want to share it, you know, as we see all over the place in the IT community, share your knowledge, help other people out. Because then you just become known as a person who's got a lot of knowledge so that when others are thinking, hey, who should we get for this? Even if you haven't raised your hand yet, they might be turning to you because they know you're, you know, really smart and you do all these really cool things. so that's actually we haven't really had and I don't know if it would really fit in the podcast but I'm sure there are other ones like how to

advance in what you want to do in your IT career from that soft skills level how do you let it be known and I'm sure there's a lot of stuff out there but it's very interesting there's those two sides raise your hand

but also don't just be quiet and sit in the back yeah it's fantastic you're do great work but people have to know about it right yeah and the cool thing at the same conference at cloud native days austria we had a last minute dropout of a speaker so on day one we announced hey if anybody in the room wants to take this opportunity 10 minute lightning talk on day number two raise your hand and let us know and I wasn't really sure if we get any feedback but also the first keynote

speaker we had on that day which is the third keynote speaker Ali Malochi he he also had a really interesting life story, refugee. He grew up as a kid in the beginning of the early days. They fled from Iran to Austria, was in a refugee camp. And a very interesting upbringing. But in the end, kind of a couple of messages where you also need to kind of, you know, don't just sit there. And he said, he looked into the audience and he said, most of you are just sitting

there and just consuming, the one that is next to you that will not just consume, but decides today to stand up and talk with more people, make more connections, and starts sharing what they have learned with others, they will in the end be more successful, in more experiences, in more connections. And this was really great what he said. And I think because of him,

and because we repeated this whole, raise your hand and try, you know, take a risk, we had seven people on day one from the audience that gave us a suggestion and what they want to talk about. And then at the end of the day, we actually put them all up and let people vote. And then we took the number one highly voted one.

So basically the audience decided what new session that they wanted to hear the next day. And yeah, it was really great. One guy, he was raising his hand, and he got picked by the audience because they wanted to learn something from him. Right, that's amazing. And, you know, that also speaks to the other side.

Who is the name of the person you said, who came over. Ali, Ali Mahochi, he was the first the opening keynote speaker. He's a known entity in Austria because of his background and what he's done

in Austria, especially he's now a UNICEF youth ambassador. His mission really is to encourage the next generation of people, not just IT, but people to understand

what opportunities are out there for them. I think even the was it Times magazine or some but not times another big big outlet they said Ali is preparing the next generation

for jobs that do not yet exist because he has a platform where he created a platform back in the days it's called watch ado which kind of like translated like what do you do he created a platform with live stories where he interviewed people around the world about what do they actually do

so that because he said when he was an immigrant and he came to Austria and when he talks with people they don't often know what is even possible, right? They only know, right, these are, I can become a cleaner, I can become X, Y, C, but they don't even know what's possible. And sometimes their background and the people around them also don't tell them or just hold them back and say, you know, be safe and do X, Y, C, right? Right, right.

Well, I think the other really, really important thing there is that, you know, everyone's got a job, you know, not everyone, but let's assume everyone's got a job, right? It may not be exactly what, you know, I want them to play drums in a band. I'm doing this right it's not what I wanted to do necessarily as a kid right but if you just do the job you're supposed to show up for

you're not going to find any join it you're not going to find any you're going to be bored you're going to be like this sucks right it's doing these other things it's raising your hand it's even just making the connections

and talking to people like having people you can talk to and colleagues that you can share ideas with makes your job so much more interesting fun and gives you invest your mental health

better into it because you're not just sitting there pushing buttons anymore you're having stimulating conversations

you're getting to know other people so I think all around that's really really fantastic advice

and it's awesome what Ami is doing I will make sure to post the link because all the keynotes

were recorded and I think by the time this session airs the recording to his keynote will also be

available really great stuff and the last thing that I remember from him he also said you know he talks to a lot

of people that tell other people you know what they should do and should not do and he said like would you say right

find find something that excites you and follow that path and he also and this is the quote that I'm assigning to him so in his closing statement

he said you are unique stay unique don't become a copy of somebody else because he was also talking about you know how we are now with AI right we can just easily generate stuff but in the

end we're just copying things we're not really you know we don't stay we're not unique anymore and he always said you are unique stay unique don't become a copy of somebody else by mimicking somebody else or using tools that just replicate things become creative and what you're passionate about yeah and raise your hand yeah yeah cool um yeah and then i was i also I wanted to quickly highlight one of the topics that I learned in Gdansk. I was in Gdansk in Poland and I did an observability meetup. And I was talking about how to analyze logs and traces. I kind of showed a little bit on, you know, kind of my career in the space and how long I've been analyzing traces.

And then, Brian, I'm not sure about you, but for me, it's always a surprise because we've been living in this trace bubble for so long at Dinah Trace. we've always done traces. It's in the name. It's in the name, yeah. And it's always surprising then when you realize that in the bigger community, in the developer community, traces are still a very new thing. And not many people are familiar with traces and how traces should be, and what a good trace can be and what it actually can do.

I'm not sure what you see. Yeah, I mean, again, I've been living in that bubble for a long time as well. As, you know, I was a customer of Dinotrace before I started. I think that was around 2009 or something I started using. Whatever Dinah Trace 3.5 was, right, whenever that we came out. Back then, well, now we call it Atmon, but yeah. So I was relying mostly on metrics back then. I didn't really have access to logs.

That probably would help. I was doing performance testing. But I know the traces were just groundbreaking, especially because we had the, you know, the auto instrumentation piece going on, or, you know, the, anyway, I don't need to get nostalgic about the old days of the Amon project.

But to me, yeah, it's so many people are comfortable with logs. But when you ask them, what's that process like? They're like, well, you know, it's kind of a pain, but I know how to navigate it. I was shocked when I saw your notes here that people still don't know quite a lot about traces, Right, again, we have been doing it for forever. The industry's been doing it for forever. The one thing I was thinking about, you know, obviously you didn't say people don't see the benefit of traces,

but if you're going to get somebody to go from a log to a trace, right, and I'm not trying to sell anything here, right? But again, the benefit of a trace over a log is logs you got to store, right? And depending on what tool you're using, that can cost a lot of money. If you need to go back, you might have to take some time to dehydrate them and all that kind of stuff, right? And it's only going to show you what's written in the log, right? Sure, if you have an exception, it'll be written there. You can see an exception in a trace. Granted, that's kind of on par.

But you're not going to have the context outside of that. But the one thing that this made me think about was as the observability world is trying to nudge people into traces. because they're more context-rich and more detail-oriented. The danger is a lot of people are going to open telemetry and the default open telemetry tracers, right, the ones you automatically put in your code,

they're just going to basically pick up the entry point and the exit point and everything in between is still going to be not seen. Platforms like ours, I'm sure many of our competitors do it as well are using other bits of technology to capture

the methods in between so you have that fuller context so I can understand the argument that traces might not be as valuable but it's only because it's like rudimentary tracing now again this is not a knock against O'Tell because O'Tell

was designed to be used deeper than the default the default is to get you started to get something in there right and then the idea being go ahead and add the O'Tel instrumentation to your critical code the idea of the third parties

adding that automatically to their code we're seeing a lot of that in the LLM world with open LLMetry I don't know is O'Tel doing something like the the snapshots yeah it's profiling profiling is coming yeah I mean profiling is already available

for Java and if I'm not mistaken maybe also they're working on dot net already so something that we've been doing for the last 10 years I think now we call it snapshots where we basically take stack trace information and then weave it into the traces

but yeah to your point so open telemetries getting there but it's interesting and I agree with your observation the default traces that just come out of open telemetry if you don't it spent the additional time to also instrument your

code a little bit more it lacks typically context because it focuses on incoming and outgoing requests and maybe some very critical libraries that are already pre-instrumented there. But they're getting there. Yeah, but I can also see too, right? You know, if your organization

didn't want to put the money your investment into observability platform, you were stuck with the logs. You've been using it for forever. So even though it's kind of a pain, you probably have all your little queries or maybe said commands or if you're using a tool, you know how to use it.

And then if you're finally getting around to it and you're just using the rudimentary hotel, you might be like, well, what is this, right? It's funny, though, because we've been, you know, even working in and using tools like Dinah Trace. What's the one Mercury had or HP Mercury? They

had that one even before. You know, it's not like the traces and stuff haven't been around for a long, long time, but as we've seen over time, it's been this growing push even in the beginning of our performance careers, getting to people to even care about performance at the first place, right?

It seemed like more of a luxury item. Now it's really on top of everyone's minds, finally. But there's still, I guess, that lag of what's all possible. And if you haven't been exposed to good traces or just haven't even had the tools that can provide them. So, yeah, I'm really glad to hear that Open Telemetry is adding profiling in because, again, that's a critical component of it. Yeah, and also, I think, big kudos to everyone in the Open Telemetry community.

who is educating. I think education is just a key thing, educating, creating tutorials, videos. Yeah, it's very important. Where did you go next, Andy, on your travels? After Gdansk, I was in Tallinn, in Estonia. Yeah? And I got to say, first of all, a funny, interesting thing. For me, Tallinn, it's the first Tallinn is the capital of Estonia.

Estonia is part of the three Baltic countries. So Estonia, Latvia, Lithuania. And for me, it's just a fun fact. I think it's the only capital that I know where you can walk from the city center to the international airport. So I actually walked from my hotel to the airport because it's like a 40-minute walk.

And it was a beautiful day. So, yeah, it's actually... I just wanted to highlight this. kudos to the Estonians but yeah I had a platform engineering meetup there and I got to say it was funny when I arrived on the

first day with an office I met the people and they were all telling me we had an internal meeting first and they were all telling me you know don't expect Estonians to be very outward like a very talkative and they will you know they're very reserved

but I typically hear this with many many countries to tell they think from themselves they're very close and then like very hard to approach. And then we had to meet up the next day in our office. And I did a quick question in the beginning, like who was actually born in Estonia and who is from remote? And I got to say a large group was from different countries. We had Azerbaijanians.

We had people from India from all over the place, really all over. But what I noticed is that the questions that I got, they're very critical or let's say they're very very to the point like very direct because I mentioned a couple of things and then they were asking

this doesn't make sense right can you know this doesn't make sense the way you explain it and and then we got into some arguments even not arguments but it was a longer discussion

heated debates heated debates which was really good but it was unexpected it was unexpected very smart people very very technically very secure and resiliency aware.

And I didn't know, for instance, that Bolt, the e-scooter and the car rental company, like the Uber equivalent is from there and many other startups in IT. I think they have the highest number of startups per capita in the world behind Israel, Estonia. So it's a really high-tech spot there. Really cool.

yeah i had um so for full transparency my mother's my mother's mother is lithuanian um but i was buying a piece of uh music gear that um i think i found it on whatever it was but it was uh manufactured lithuania and i was oh wow you know i had you know and this isn't to knock anything about that region but like it just never comes up in anything and then i've seen some since, like there's more, you know, gear products coming out from that area and all. It's like, oh, okay, there's a lot going on there. And as you're saying, they've got, you know, a lot of, I didn't know, I didn't know idea they had all these startups going on there.

That's pretty amazing. Yeah. Yeah. But it's always, it's always interesting once you live long enough to see different areas of the world pop up, becoming more center stage and other ones falling back a little bit more or, you know, taking a back seat. I mean, it's obviously also a little bit. of a of a bubbled view because you know the reason why we talk about is now because i was

fortunate enough to travel there and now it comes up and i'm sure then you are you're more receptive to stories like you with lithuania right maybe you have if you would have been able to see it but you didn't have kind of any connection to it and once you found out about your uh the mother of your mother right the grandmother i mean it's like then things all of a sudden become more visible yeah yeah cool Yeah, and I just wanted to say, thank you so much, everyone that I was able to meet in those cities, in those events. Thank you so much for the hospitality, for the good conversations. And hopefully, hopefully there was at least one or two things that I could help people with inspiration of new thoughts and sharing some of the stuff that I've learned, and say it that way.

Yeah. You don't share stuff that you learn at all, though, so. sarcasm. It's awesome. I always get jealous about how much you get to travel and meet all these. But again, that's going back to Ali, right? It's travel around and meet and experience all this stuff. Anyhow, so what's up next then? What was up next? Yeah, what's up next? Up next is a little tour that I would like to promote, Henrik, our dear friend. I've invited him to also be part of today's podcast. Quite a long time. Hi, Henrik. Yeah, he, I tried to make him also come today but he's on site with a customer right now traveling and he and i we go on a little tour

in the u.s we're both at cubecon in atlanta so folks if you're listening in we will be in Atlanta obviously doing cubecon and cloud native con and all the co-located events but then henrik is traveling to the u.s a week prior and i will be staying a week longer and together we're doing a couple of cities where we're doing meetups and and events I'll post a link also in the description. I think Henrik is doing Chicago, Cincinnati, Dallas. Together we'll be doing Atlanta. So we're doing a Dinah Trey's event in Atlanta.

And then the week after, I will be doing Omaha, Minneapolis, and Boston. So it would be great to meet some of you. It's open events. So that's why I just, I'll send the links. But it will be nice, especially. Yeah. Yeah, especially the topics right, Henrik will cover a lot around open telemetry and cloud native, Kubernetes, that is his world. And I'll try to talk about platform engineering and stuff that I see how AI can help.

But also, I think I learned a lot where we also want to be a little bit cautious, not everything needs the AI sledgehammer. There are certain things that can be solved with simpler and more cost efficient and more deterministic solutions. But AI is obviously here to stay and we need to just figure out how to best leverage it in our day-to-day life. Right. Well, I also can't wait to hear what you come up with some of the AI-native stuff.

I was talking with my neighbor who's got a small startup AI company and, you know, I brought up the topic with him and he was telling me how like, well, you know, they started building from the ground out. Like, again, he said they were AI native, right? But again, as we know, there's no full definition an AI native, but what he was clarifying from that was that they weren't thinking about,

I forget how he put it, but basically they considered AI about everything because it was going to be an AI platform. So it was more about what do we need to accomplish. Now let's go back and find what we need to use to build that. So it wasn't this, let's, you know, it wasn't the lift and shift to the cloud, right? Or lift and shift to AI. It was a brand new thing. So they got to really think about it strongly beforehand and try to take into all that kind of stuff into consideration.

But obviously, this is still a burgeoning topic. So I'm really looking forward to either some guests we can pull in about AI Native or anything you might have learned from that bit as well as, you know, what's going on in Cloud Native. Well, we're talking about guests. Also, we will have Pino Riesnick. hopefully they pronounce his name correctly he will be back because his AI native book

has been published I think two weeks ago and he's coming back in November on a peer performance episode you know one last thing I was thinking of just as you were talking about doing the tour if we had a bigger podcast I don't know if you ever seen like these really big podcasts you know a lot of usually they're like comedy ones

or whatever but they'll go doing like tours around the country recording live podcasts and theaters and stuff like we've got a bigger podcast Andy we could do it to hey everyone I think I'm to listen to us

yeah yeah I'm comfortable yeah I'm gonna say I'm comfortable with the setup that we have right now but who knows definitely we'll see all right yeah maybe the last thing I wanted to quickly scratch upon because we've been

because we just talked about the AI Native book some of you may have followed some of the postings that I did last year we published a book on platform engineering so it was called platform engineering for architects together with Max yeah you got it here yeah did you read it

yes yes just give through it every time if I pick it up I'm like oh let me take a look at some oh that one looks at some pictures I haven't read it you know front of that's fine So, yeah, after this book, right, I thought I'm done with writing a book.

But then Hillary, one of my co-authors from the first book, she said, she would love to do another one. And we got in touch with the publisher, and they had the idea to talk about modern observability for cloud native and AI. So like kind of AI ops, what is it really like now in the era of cloud native and AI? So how can AI benefit observability? What are the use cases? What does modern observability even look like in the cloud native, in an AI native world? And so we've been starting to write a couple of chapters and have also a third author, not Max this time because he didn't, he couldn't do it.

But we have Rob or Rob Rattie, who is with us that Hillary knew from her past life. And yeah, looking forward to that. It's been an interesting journey again, a lot of learnings, a lot of time spent in writing, but good. Let me ask you, when you talk about AI ops, I see that thrown around, and I'm sure our listeners hear that a lot. Are you speaking from the point of view of how do we observe AI platforms, or are you talking in the point of view, how do we incorporate AI into our observability platforms to make them more efficient, more powerful? Yeah, I would say more the latter. And I think the way I started, start Chapter 1, I think Gardner coined the term AIOps 10 years ago, right? And AIOps 10 years ago was all about using some machine learning and then basically denoise all of the different events from different separated tools.

This was the time of the MOOCs softs of the world. I'm sure there's still around. But, you know, you collect a lot of different pieces of data. and then there's so much noise and you try to use some type of algorithm some type of machine learning to denoise to make your operations more efficient and I think this was one of the initial definitions

from Gardner and we are trying to say you know kind of 10 years later what's the world what is modern observability and really do we have better ways than just denoicing how can we enrich observability data

from the beginning with more context this is where standards like open telemetry come in where the semantic dictionary come in. So instead of having multiple separate data silos and event generators, we tried to enrich the data at the source with enough context, connecting logs, metrics, and traces in our profiles and real user data and whatever else. And then use modern algorithms to make sense of this data.

So that, you know, optimizing the work of a human when it comes to detecting anomalies, finding root causes. This is obviously stuff where I bring in some of my background from the work that we do here at Dinah Trace. But the book, we write it in an independent, right, a thought leadership way. So everything we write there is meant to be just best practices that we have seen out there, how you would do, how you would implement observability, the challenges that come with it. And then obviously, AI has a big part of it. But we are also very critical.

And we say you don't need AI for everything. It's indeterministic. There are certain things where simple automation with a couple of scripts that get you the data that they need in a consistent way. That's enough, right? You don't need to throw. You don't have AI write those scripts for you.

You could have AI write those scripts and then you automate it. And then you get a deterministic answer at a reasonable cost factor. Yeah, I think the next 10. years of tooling and seeing how all this stuff is incorporated, both in terms of, as you say, denoising and all that other kind of stuff, but also the AI assistance in helping you create and generate the automations or dashboards or whatever else you need to do are going to be very, very interesting, especially once we figure out how to start connecting different tools together. Obviously, whenever you're talking about connecting data sources together, you have to worry about. security, data privacy, and all that kind of stuff. So once we, as an ID community, figure out the best ways to do that,

to be able to leverage all the information out there in context of, you know, if we're talking about observability, my company's platform that I'm observing, right? It's, you know, the possibilities are really tremendous. And as we've already seen in this AI world, things develop really fast. There might be, you know, a little chunk of stagnation. And then suddenly, bam, we have a tremendous leap forward. So it's definitely going to be really interesting to see where this all goes in the near future. Yeah, so I think that's a little bit of the update that I wanted to give you.

And I'm very much looking forward to all the trips that are coming up, and especially to our upcoming guests, whether it's Laura, whether it's Pini. I have a couple of other guests, some of the speakers that I met at Cloud Native Days, Austria, and then a very old friend that I've known for years just reached out on LinkedIn, William Love. We will have an interesting conversation, I'm sure, as well, in December. And yeah.

Awesome. Well, Andy, I appreciate you being a guest on the show today. No, seriously, this is always fun. to have these kind of updates I think there's just a lot that you get to see and do and you know as you mentioned earlier

from your time at where was it Austria Cloud Native Day Austria where Ali was talking about like it was the right you know

sharing information right is really key and it helps people know not only learn but also know what you know and be like oh Andy is really smart

I want to contact him for whatever, right? But really, it's just always a testament of your willingness to share. And we see it all throughout the IT industry. I mean, the whole basis of open telemetry, right?

It's not open telemetry. Open source, right? It's sharing and helping. So it's alive and well. Thanks to people like you out there, spreading the words. Thank you, Eddie. Appreciate it. Thank you. I hope to have you back on sometime soon. Yeah.

And thank you to all of our listeners. Any final thoughts there, Andy? Or you're all good. No, I think I'm all good. I'm just going to close with, raise your hand. Yes. You know, tell the people around you what you are, what you are capable of doing,

what you are willing to share. Yeah. And if you find out something cool, share with other people because chances are they don't know about it. And that then helps you share the knowledge but also people will think, oh, this person's pretty in tune. I should pay attention to them, right? Or at least, you know, your management or whatever.

Yeah, both of those are fantastic ideas. And, yeah, thanks everyone for listening. We'll see you the next episode. Bye-bye. Bye-bye.