Andreas Prins, StackState | KubeCon + CloudNativeCon Europe 2023
With the adoption of Kubernetes rapidly accelerating, many companies struggle with having the right skills within development teams to troubleshoot incidents quickly. Remediation of issues is of the greatest importance to avoid customer disruption. This interview introduces several best practices where SREs can take a leadership role, such as:
Knowledge sharing – and the importance of it – to increase developer expertise
Enabling SRE teams to take a leadership role in coaching development teams
Encoding expert practices in easy to understand ways, to help novices learn
Implementing the process automation necessary to ensure high quality and detailed insights
Transcript
This is Textron TV. Hey everyone back here in Amsterdam for our live Cube con coverage. We're wrapping up day three.
I can't believe three days went by this quick. It's been a blur but we've had had some great conversations. It's 10,000 people here.
The buzz is it's back right speaking of back. Let me welcome back my friend Andreas. That's correct.
Absolutely, right? All right. I got it right Andreas from stacked stated.
It's funny. We were talking here off camera. The first time I interviewed on jayas was in nice at a devops world.
You were exhibia back then Jimmy laughs definitely and You know the world goes round and round it seems doesn't it? And and here we are now Andreas, of course as I said, it's with stack state, but on JS besides working exhibia Labs. Why don't you introduce yourself to our audience now?
Thank you. So Andres Prince and the born and race here in the Netherlands. And I'm working for steak.
Yeah, definitely realize that this is my home game. Absolutely homeboy local boy makes good, which is nice because my entire engineering team is here as well. So it's a kind of our social outing being together learning a lot good guys, and I'm leading the the product team at stackstate.
So responsible for the strategy the direction we're taking and super exciting because the market is pivoting rapidly. So let's talk about it a little bit. Absolutely later.
Yeah, we're gonna talk about the market, but first I wanted Talk a little stack State not everyone here is familiar, you know in our audience. Yeah. What would you tell them?
Like? How do you explain Stats today is actually really simple all of you know, grafana data dog who bring you a lot of dashboards and that's helpful because then you have insights on what's happening, but you don't have the answer yet on how to solve if issues arise in the landscape. There's Tech state does observe ability, but the next step they do is they pick up a signal and they attach what we call a remediation guide or a Playbook if you like to it.
So if you suffer developer can then troubleshoot the issues that arise in his cluster and that in particular is important because The kubernetes layer we've seen that here around the kubecon is more or less. Well established. There are many tools the ecosystem is maturing, but now a lot of software development teams start to deploy to the kubernetes cluster and then they run into trouble and then the question is how do I remediate the issues that arise and that's where stackstate at its unique capabilities to help regular teams out and that is important because they want to focus on software development and not on knowing kubernetes Inside Out detailing or knowing all the details there, you know, two important things you spoke about I want to I want to head on first of all, you're right we've seen this Evolution.
I talked about it with someone earlier from Dinah Trace. Yeah, 17 years their field CTR we've seen this evolution of what we do with this data over the years. We've called it APM.
Yeah. Just playing monitoring APM. We you know, we now Call it observability.
We've called it AI Ops ml lops. We called it a lot of names but it's still underneath. It was always the same we wanted to.
Get the metrics to somehow instrument. How our apps how our infrastructure how it was performing at first. We were happy to just be able to collect data.
Exactly. Yeah, then we started collecting data really good. And we said, oh my God, we got so much data.
We don't know what to do with it. And it's expensive to collect on Yeah. Well, yeah.
Well, maybe we shouldn't collect everything then we started saying. Oh, yeah. Well, what should we keep it?
What then why do we do it? Yeah, how do we yeah analyze this data. Well, we'll try ml.
We'll try AI we try this we try that. Now we got this new term observability right and now we could get some actionable intelligence. on data But now okay.
So now we know there's a problem but between you and I Andreas we knew there was a problem. Yeah, because I'm glad you calling out. So what we so if you take a look here at kubecon, there are a couple players.
Yeah, let's say kronosphere CoreLogic who understand the data or too much data becomes an issue. So you see Solutions now and you observability space who really focus on can I reduce the amount of data more accurately at the intake of the data? That's one Trend we see so okay.
There's too much data how to how to lower it. The other thing is the awareness of what to do with the data. So if you have done the data, how do I make it actionable and we've seen the move of slis slos yes or more pushy to building an understanding.
What do I need to do from a business perspective and funny enough. We thought we were super unique with our capability and I still believe we are with our four more players who start to utilize the data for the purpose of troubleshooting and that to me is the first conference where we meet other companies who also start to realize A dashboard is nice. But how do I use the dashboard for the purpose of getting back in a healthy stay good?
Yes quick as possible. Yeah, and and that and that really is stacked State Special Sauce, right? Yes, you grab the day you could use any tool to grab that.
You could store the data. You can analyze data. What do I got to do to make it right?
Exactly exactly. I like the way you so the second thing I wanted to mention is I like the way you use the word Playbook. Right because whether it's a run book A playbook a recipe.
You know puppet script. This is something devops understand exactly right that we have a repeatable. Script.
Yeah repeatable pattern. Yeah of how we we, you know, do these things. Talk to me about that a little bit more.
So the The Playbook I suggest is not in us. So if teams start to do post mortems, so an incident has happened and they want to analyze the incident and I really want to know hey, how can I avoid this happening again? They create a Playbook a list of actions that whenever it occurs again, you can really quickly solve it what we thought is but That's after the fact.
How do we bring it really early on in the process or as soon as you see early signals of it might feel how can you then bring the Playbook into the hands of the developer so he can himself check the locks the events but really guided step by step if this issue occurs, then this is your playbook that you need to follow. So rather than and I would say it kind of Or not evolution of area office was just almost magic, right? It happens in the books with AI and ml.
But what is happening? You don't know and now we made it more simple we say okay, there's a signal and then there's a remediation guide that you need to follow and that is in particular helpful because what we hear over and over if you want to roll out cloud and kubernetes platforms to bigger teams in the Enterprise. They're only five six people in the platform team or as a team who really have the knowledge, but what about all your engineers?
And these playbooks are a really good instrument where the SRE or the platform engineer can kind of encode his knowledge do that once and then for every situation that it occurs. Well, then the tool will provide the Playbook to the engineer. So it's also a really good way to capture the knowledge and to share that in your entire organization.
Very absolutely. If you don't mind, I'm gonna pivot a little bit say, you know, you've had a lot of conversations here over the last three days. What are you hearing or seeing that kind of?
You want to share with the audience that you see is maybe trends that are coming out here. Yeah, you know emerging. So not exactly close to to this next State's business but slightly broader the rise of the developer portals or the developer platforms.
If you like, it's they're quite a big of quite a bit players around here that slowly start to understand okay with agile and devops we've gone really wide freedom for every engineer and now they realize okay, there's what's probably a little bit too much Freedom. Everyone has their own tools. How can I bring that back and really focus on onboarding developers real fast with the right tool suite and even beyond that tool street, so that's that's one and the fact that all of a sudden our three four five of these companies at the single conference to me shows that that market.
Yeah, he's maturing. It's a sign pretty rapidly. So that's one.
Well, obviously at the platform management, it's cool to have kubernetes but how to do when it becomes real the third one. Security everywhere and no no dots there right in the middle right front and center but that was already previous two corn and the coupon before as well. So that's that's a continuous a continuous stream.
Yeah. Yeah, and then what I mentioned I see that pivoting observability. Yeah.
Okay. The dashboard is nice. But how do we make it more more actionable?
So I would call out these as the kind of the four important ones. At the same time and that is there's not a critique to anyone. If you join sessions, it feels to me sometimes that they're super basic.
And that is at the same time to me a signal that the market is still in its early days maturing towards yeah. The mass is head towards the the bigger adoption and there was a surprise to me. So I was expecting a lot more talks that are also deep talks.
But also a lot of talks where about the foundations of observability or monitoring and obviously kind of a surprise that I would think hey people would know that by now. We're in this for five six seven years. It's real already for a couple years, but still there's yeah obviously a big market for that.
So, you know, what because we live in bubbles. And sometimes you know you go around to and and it's also this whole like crossing the chasm, you know. Yeah.
We're there. We're exactly at that moment. Exactly.
You know what you're you're dealing primarily with early adopters those 15% but then it crosses in the mainstream and they don't adopted so quickly. So observability and some of this stuff is it's just coming. It's just nothing.
Yeah. Yeah, you know it I I think the growth curve is still ahead of us up behind I think so as well. That was that was my lesson stock still definitely.
Oh super happy to know that was my lesson as well. Absolutely. Yeah, and then one funny we started the prior to the recording about release Automation and cicd.
The funny fact that's an established Market. I don't see many new entrants at The Usual Suspects are here the harness the coach. And that is obvious because that market is about 12 years 14 years old.
So that is that's a mature Market. Well still growing only thing it's still going but I'll tell you the get-ops because to me get officers where cicd lives now. Yeah.
Yeah, that's true. You know that there's some stuff going on there. You know, I interviewed a few people.
It's very good. But we'll see I think the more you know, that that we'll see how that develops because I'm not expecting a lot of new entrance. No in this people are maturing they're offering and more people are adopting.
I think Ai and automation still can twist and shake that market a bit, but we'll see anyway. Hey, we're almost out of time, but I want to For people out here want to find out more about stack State. What's the on-ramp?
What's the easiest way for them to get engaged? com and we've provided a really great playground for you. They're even a couple kubecon challenges where we've introduced issues in the cluster as Tech State takes you through and you can step by step remediate or not remediate but detect what the issue is and then if it would be your own collage, you can remediate it.
And that's really the easiest way to to get started. I love it Andreas. It's a pleasure.
Thank you so much. Go check out stocks today. What was it called?
com coming on here. We're gonna take a quick break bring in our last guest a few con stay tuned.





