Techstrong Gang – April 3, 2025

Hey everyone. I'm Alan Shimo. I'm a cube con.

You are not. You're watching Textron Gang. Well, maybe you were a cube con and you're watching this after you were there.

If so, I, I hope you enjoyed the show. We're here down in our, uh, cube Con booth and two 40 too late. By the time you watch this, we'll be gone.

But we're in N two 40, uh, having a great time with 12,000 odd other cloud native aficionados. Welcome to our Textron Gang Coverage of CubeCon. I'm joined by my two CubeCon, uh, through your cohort.

Co cohort. We're a cohort. No, we're co-host.

Well, not too. Okay. Mitchell says we're cohorts.

I think cohorts have a different meaning. Oh, okay. So they got, you know, well, we are kind, there's a shady part of this.

Yeah. There is some sketchy there about, anyway, they're my co-hosts as well as cohorts. I was gonna go with partners in crime, but Okay.

Alex sketchy that better. Anyway, to my extreme left. He's not really extreme.

Well, maybe. Um, future of Alice, Mitch Ashley. Hey, Mitchell.

How are you? Hey. Good, good, good.

Got and the man in the middle. Yeah. Okay.

And then we've got the man in the middle. He's not gonna do a man in the middle attacking. I hope not.

Hope not. It's Chief Content Officer Mike Ard. Hey, Mike.

How are you? Isn't There a song about something in my right of me and something in the left of me or something Stuck in the middle? Back in the middle of you?

Yeah. Yeah. Something like that.

I thought you were gonna say Riding in the Desert with a Horse With No Nick, but there you go. That was just a song I had buzzing around. Anyway, we've been, we've been, we've been a Q con too long.

It's only day one. So guys, what do you think? Well, the conference is huge and I've only been here actually one day and I swear I've been lost five times, So.

Well, the, the venue here is the Excel. Yeah. It blew me away yesterday when I was walking around, like trying to get my bearings.

You know, I have a suggestion though, for next year. I didn't say she just park the planes on the side. And this can be a con horse and airport all 'cause there's an airport real close by, right.

I'm flying out of that airport. Yeah, I mean, save us that train ride. That would be nice.

It is. I mean, what's interesting is, look, I gotta give London a lot of credit. You, you know, they've rebuilt, we went last night to dinner at Cannery Wharf, which is a financial center now, but mixed use residential and stuff real nice.

This area, they're building like mad. You know, you're a good way away from Central London, but mm-hmm. It's still, this is a kin when in New York, remember they started building Astoria in Long Island City and Yeah.

This Brooklyn, this is A, this is almost like if they built skyscrapers in Howard Beach outside of JFK. Yeah, exactly. Not Dumbo or, no, it's not Dumbo.

Not Dumbo. But, um, anyway, yes, it is a very big venue. There's two main rooms for the ex exhibits.

We're in the north one. I haven't been over to the south yet. Uh, you can get lost.

I like that they have food restaurants in the concourse too, though. Fire places can sit and chat and have Yes, there is. That is awesome.

They've done a nice job. But anyway, let's talk about what are the big stories, Mike? Give us our three, uh, Well Spots here.

Let's lead off with what Microsoft was talking about in the keynote where they have a project they've been working on quietly called Headlamp, and it's a graphical environment for Kubernetes. And now apparently that's gonna be given to the main Kubernetes open source project team and integrated into Kubernetes. And they're saying that this is how the next 10 million users of Kubernetes come to, to be, because they're gonna use a familiar graphical tool.

They may not have to have as much DevOps expertise to do that and take a little bit while before that actually comes together. But, um, on the one hand, I listened to that and I was thrilled. And on the other hand I was like, what the hell took so long?

Mitch, would you like to go first on this? Another dashboard? Oh, good.

Yeah. Great. Look, that's my first reaction.

It's like, okay, let's talk about how that fits into the work that happens. Is it, is it really a dashboard or it's windows sizing? Either Windows izing and is really, it's a full graphic.

I made up a word in your copy. There You go. I like this.

I feel Like Ralph crammed in and Norton. I, I would say, you know, say what, say what you will about Microsoft over the years. The one thing that they do really well is great graphical environments for things especially, I mean, you look at Windows, it's a, historically it's Back rip off.

I'm only kidding you, But it's always been, you know, a well designed system for dealing with something that's in lower level of abstraction. So I have a few thoughts. Number one, anything to make Kubernetes easier is a good thing to manage, you know, to manage the Kubernetes, uh, infrastructure.

It's a good thing and I applaud Microsoft for doing it. I'm a little unsure though. Headwinds, Headlamp.

Headlamp. Excuse me. I thought we were running into headwinds cohorts.

You, you're looking cohorts. Anyway. Headlamp, is it its own project?

Is this number Project 203 for CNCF? No, that's, or is it just part of the main cube? It's going right into the main cube project.

So it's like a package for Cube. Similar like Linux packages. Yeah.

I don't even know if it's a package yet. I think they're still sorting out how it goes into Kubernetes, but I think it's gonna be core into the framework and not some one of 347 other projects that are running around the CNCF. That one seems to Remember.

So here's the question about it. Are they, is this a apply to get more Microsoft customers using Kubernetes on Azure? Right.

Even though it's open source or, you know, Microsoft has a very distinct look and feel in how it operates with its user interface, which is probably not gonna be appealing to the latest Linux admin. Right? Oh, contr.

Oh, oh, really? When did you become a Linux admin? This is the year for Linux on the desktop.

No, But uh, But that being said, that being said, the fact of the matter is Azure, on Azure, it actually runs Linux predominantly. Oh, yeah. I mean, 365 runs on Kubernetes.

I mean it's, Yeah. Yeah. I mean, it, it, so they, my point was, Is this really, because it's gonna be very familiar to Microsoft customers, the ui, will it make it even easier for those customers to More I think it will.

I think it will. And I think you can't underestimate how many Windows users there are. They already voted with their feet not to use Linux.

So they clearly want something else. Yeah, I agree. But how many people would like a gooey for coup?

I think the issue is there's not enough engineers. And so if I need people and I'm gonna have thousands and thousands of clusters, how am I gonna manage that if I don't make it accessible to mere mortal IT administrators? And that's the only way we're gonna get there.

Get some graphical, I Think it's an evil quad to get more people on Azure. Yeah. Yeah.

I don't this, it doesn't have to necessarily be an evil, but it, it gives Microsoft customers an advantage. My my point being, if you're already familiar with all the, uh, operational tools for Microsoft environment, now here's a EY for Kubernetes that is gonna look really familiar to you because you already know how we operate. Other folks are gonna say, yeah, we're not a Microsoft shop.

We're not used to, we're a gruff. We're this, uh, that or we, whatever. Okay, yeah, let me kind of use it, but it isn't just the natural plug and fit.

I'm ready to use it. So that'd be my point. So there's more at work here though, right?

So Wait, there's more, There's always more Order now. So Reshipping and hand, At some point that graphical environment is gonna have AI agents all over it. Yeah.

Yeah. And what that looks like is gonna be interesting to see, because right now, I don't know if there's gonna be like a, a super agent that manages all the other agents, like, you know, the head butler from Downton Abbey who managed all the servants, or is there gonna be like, you know, son of what was the Microsoft thing where they had, uh, clippy, clippy, son of clippy and there's gonna be like all these agents all over popping up saying, oh, can I help you with that? Or can you know my turn to do that?

And at some point the noise level gets too high. But I think what's gonna happen is a lot of these tasks that we used to have a, an administrator handle will be handled by an agent that's being supervised by the administrator, assuming there aren't too many annoying agents. Mm-hmm.

How do you see this playing out? Well, Again, I, I think that's a another positive Microsoft direction because as you're developing agents in a, in an Azure, Microsoft environment that fits right into this open source project, the gui, the plugin, the APIs, whatever you're using, I think it gives them an, an advantage. Now whether that drifts and becomes a little less Microsoft centric as an open source project, probably will.

But I don't, I don't, I haven't taken a look at headlamp yet to see if it's kind of thinking about agents and done That far. I don't, I don't think they're there yet, but I think that that's probably coming. Next question is, will people who have grown up, uh, managing Kubernetes using CLI and YAML files, give that up in favor of a graphical Approach?

Oh, wait, let me think about that. No, I would do it in a heartbeat. What, are you kidding me?

Who wants to play with YAML files like that? They're not gonna give that. Yes, they are, but let me tell you where you got off the tracks here.

Wasabi, mark my words. Agentic AI and, and even generator ai, AI marks the death No. For graphical windows based interfaces.

Because even the greatest Microsoft users don't love the whole mouse window paradigm. I want to, I wanna be Scotty, hello computer. That's what I want.

And that's what people want. They don't wanna maximize, minimize mouse this, drag that they just wanna tell the computer what to do and the computer does it. Let, Let me take the same place, but from a different, for a different reason.

Okay. I think that the graphical UIs Yeah, if you don't like command line, yes, they're definitely better. Large volumes, large amount of functionality, large, not data is still really difficult to do in a, in a graphical user interface.

And to your point, like, I wanna bubble up the supervised tasks that I would need the agent to run in collaboration with me, or knowing what triage it's already done so I can take the next step on this problem. It's kind of bubble to the surface in a ui. Not now I can go search 8 million log records in one place across all my Kubernetes clusters.

I, I, I think you're right about AgTech taking this over, but I think it's the end of the gui. I I think we'll know those boomers by their cli So you're saying, uh, graphical user interface, it's the acoustic coupler of user interfaces. No, I'm, I'm saying modes by the way, if you don't know what that is.

Oh, no. But let, let's look at this, right? Wasn't Xerox Park 1968?

71, 77, yeah. 71, 72. So it's been 25, it's been 50 something years since the, the paradigm of mouse and windows became sort of like, you know, the go-to way of interacting if you didn't want to use just a pure CLI.

And I think people want and deserve better. We deserve better, and we are going to get better. I think agen ai, generative AI chat interfaces, better human machine interfaces are finally capable.

Yeah, I think it's natural language is the ui. Yeah. And in the future, the natural language, whether it's chat bot or interface to any app.

So that's, that's my take. And I'm going, I'd like to event then and call it headlight On, on behalf of the entire IT industry, I'd like to apologize for subjecting you to user interfaces and CLI for the last 50 years. Years.

Well, 10 ha Marys. And as they say, people vote with their fingers and feet wallets quick. So we'll see where that goes.

Um, alright, moving right along. We're gonna take a break here at Q con. We've got our next block to go over.

Mike, you want to give him a teaser? Google. Google.

Don't, don't Google that. Don't Google. We'll be right back with more.

You are watching techron gang. Discover Techron Group, the epicenter of tech innovation. We are your go-to for reaching IT leaders and practitioners worldwide.

Our secret impactful content that sparks awareness, engagement, and top quality leads with us. You'll access editorial websites, streaming videos, virtual events, custom content analyst research and more. Join our satisfied clients.

Let's revolutionize your tech journey. Contact us today and tell your story to the world in the most powerful way with Textron Group. All right, folks, we're back.

And yes, we're talking about what Google's doing here at the show. They are, it's a public preview, so it's not quite ready for everybody. But they've got a new orchestration service that helps you manage fleets of Kubernetes clusters and you can have more granular control over how infrastructure is allocated.

And this too is another tool that I would say is long overdue. But what's interesting is that one of the reasons that they're so keen to do this is that they apparently are running outta infrastructure and they don't have enough data center capacity. And so now they're investing time and effort to make it organizations consume stuff more efficiently.

Heavens to Betsy. What a great idea. So he didn't believe the mi he thought Microsoft had mal intentions.

You, you swallowed That. I'm just saying that, you know, I don't think they would've said that three years ago. I think now they're like, holy moly, we got an issue here.

We can, we don't have enough capacity for all these people. Well, I'm glad to see they admitted that at least. What do you think, Mitch?

I Sat next to a gentleman from a very large financial institution, platform engineer, head of, of their engineering who are running 12,000 Kubernetes cluster. This is a global, right. Global.

Oh my God. How do you do that? So people are solving this some other way now, he didn't say, I can't wait till somebody introduces another open source project for me to use.

So I, you know, I asked myself, why would Google do this? Right? Is it, is it to increased adoption of Kubernetes?

Probably not. It's increased adoption of Google Cloud services. Yes.

Right? Bring it into our environment. Again, it's the same, it's the, yes, you can use it wherever, but we've got it built into our fabric as services and tools is all part of our cloud.

Right? Scale it up as much as you want. Uh, to me that's the motivation.

So I'm, oh, go ahead. So I come back to this, uh, full circle, just last, same as the last conversation. They will inevitably put some AI agents in this thing.

And so we're gonna be able to automate the management of Kubernetes clusters at scale because Geminis, yeah, a little Gemini, a little shake and bake. And then you throw it out there. And I'm not sure why this is even a separate service.

This is more like a feature to me of when I buy the Kubernetes thing in the first place. I wanna appeal to our audience here. Let me, let me ask you folks.

Are you sensing a pattern with my cohorts? Mitchell here thinks everyone does everything so they can sell more cloud services. Michael says it's only an intermediate step on the way to agent AI Nirvana.

I think there's more to this world than those two things, really. Yeah. What else is there?

I think Mitchell, yes, some of the biggest companies in the world can manage the thousands of Kubernete clusters, but I don't think mere mortals can No, no, no. I'm very, and and I do think we need, I do think we need, um, solutions for the rest of us. That being said, yes, companies do things that are in their self-interest, obviously, but if what they're pushing isn't in the best interest of users or it doesn't have an appeal for users, it doesn't get adopted.

Ultimately the market is a crew master and it rewards things that work or are good. Even if the underlying provider may, you know, benefits from it. Yeah.

I think that they were serious about, you know, getting people to consume infrastructure more efficiently. There would be some sort of built in finops capability that I can see what my costs really are in real time and then I can make intelligent decisions and then I could use this new service to dynamically scale things up and down in importance with a cost meter that would let me figure out what I'm actually spending at a given moment. 'cause honestly, I don't think engineers and developers go out of their way to spend more than they need to.

There's just nothing there for them to look at that says, wow, you know, you could think about this differently. And so then they just, you know, turn it up to 11 because it's like availability. And I don't want to get a call at two o'clock in the morning.

Connect The dots. If you're doing that through what Google provides, now you're aggregating all that data around a Kubernetes environment. You can now apply ai other things to it to say, Hey, let's auto size this, or let's recommend some ways to save money you on services or make your apps run more efficient or resize the memory footprint of containers, whatever it might be.

You've got all that that you can apply. You've got all that data sitting right there in Google Cloud to do that with, with ai. Sounds like another evil plot to sell more Google Cloud.

I don't know, I'll say what I said last time in the earlier segment. What the hell it took so long. This has been a problem forever.

Have You guys been drinking? We just on riff and repeat Keep. Yeah, you're, uh, I feel like I do.

I press button A or button B. This is two TV channel. Okay.

Network. Maybe, Maybe it just reflects that Microsoft, Google and everybody else who's in that cloud business, the differentiation levels are zero. Well that's true.

I, and I think, and it's interesting, we're actually gonna be interviewing AWS here in a bit, and I'm interesting to see what they come out with to sell more AWS Clouds and eventually introduce agents and what took 'em so long to do it. Um, now, You know, the three Questions to ask. They, they pretty much to, to how's it gonna sell more Clouds, sir, to, To their credit, they just did integrate Amazon queue developer with Amazon Open search so that you could do agents for incident management and they're, they're on there.

They're, that's the second piece they did one last year where they have an agent that kinda, uh, does some analysis through Q developer. But you know, they're stitching all this stuff together through all these different existing services. And I'm kinda like, blow up the services already.

You know, people need like a capability and an outcome, not another thing to buy Or give the agent your credit card number and let it get what it means. Yeah. But let me ask you a question about this whole clock for world domination that you guys are on here.

What about people who do multi-cloud? I I don't wanna buy a Google only solution or a Microsoft only solution or, or even an AWS only solution. I don't want to be locked in.

Yeah, there are multiple third party platforms out there that do a lot of the same things. The only difference with Google is they're gonna say you already have an enterprise agreement with us. So, you know, you can use your credits for that to turn on the Google service and you know, we're only gearing to it as a service.

So we can keep track of how many credits you consume, but otherwise I gotta go buy a platform that's separate, install it, manage it. Same with your finops, right? Same with my finops.

Now, you know, is there gonna be an AI agent framework that sits above all these things and just sucks data out of these things? And then I, and they're all backend services that I don't have to deal with in the future? Maybe, but I don't know who's gonna step up and create that heterogeneous AI agent for IT management.

But maybe it's a KY Jewell or some consulting company. Well, there's also the, that the underlying thing you're managing is open source. Why couldn't I use the Google tool to manage my Kubernetes running in AWS or, or, uh, uh, as, as you for that matter.

I mean, maybe I don't have all the capabilities in that native environment, but I could still run it from my console, my whatever, from native management orchestration tool operating out of Google Spot. I think it screams for an open source product that does this across multi-cloud and private cloud. My god.

Another CNCF project. Yeah. What are You crazy?

Let's talk about that. How many CNCF projects are there? Uh, Lemme submit this one, right?

Yeah. More, more than more star. There are more projects than stars in the sky.

Like Sands in the hourglass Sands of time. I mean, honestly, how many can you, can you have at one coupon? How many can you actively manage and manage?

Well, That's a question they'll be asking forever. How many of those projects are gonna here but orphans because they're just kinda hanging out on the edges of the thing? No, I I think it's, it's something different.

I think. Think of baby birds in a nest. Oh, you're gonna kick the, the weaker ones out.

Well, if the one bird gets all the worms that mom and dad bring, that one bird grows much bigger and then commits, uh, sibling side or whatever the word is, killing your siblings. This Explains Big bird Kind of. Yeah.

Big Bird was once had brothers and sisters. They were all medium birds. They left that part outta Sesame Street.

They did. Well that was, that was cut when the sensors came in. It was not rated G.

But seriously, is this inhibiting the growth of CDF Open Mainframe OSSF everything except PyTorch Open Telemetry Open. Well, no. Open Telemetry is part of yeah.

CNCF, but our other foundation Oh, I see. Other foundation. So because all the projects come here.

So in theory, the fact that all these projects are under the same, uh, consortium is supposed to promote cross pollination of contributions across that, but it doesn't seem to be actually happening. And No, some of them are directly competitive, but how many mes down That's in the market spoken. We need a whole bunch of different consortiums underneath.

Or we can aggregate it into CNCF. Do we really need these other work? Do we need A-C-N-C-F If you have a Linux Foundation Yeah.

Or they announced today, we did an interview on it. If you wanna watch Open Infra. Yeah.

Open Infra Foundation is now becoming a top level domain, a top level foundation, right. With lf that's, you know, the o uh, OpenStack and stuff. So Consortium consolidation, it's coming next, right?

You Need that or, or not? Or or do you, does every open project, open source project deserve its own foundation? Its own infrastructure.

Its own everything. Well, That seems a lot excessive redundant because you got lawyers for each one and there's all marketing PR people, there's all this backend stuff that these projects need. So that should be kind of a shared service.

And it's, in my mind, it's a question of how do I get the engineering talent to kinda work on these things more collaboratively so that they are consumed more easily? In fact, one of the things that was telling about service meshes and the latest CNCF research is out and the percentage of people using a service mesh actually dropped because, and the assumption is that it's just too hard. Mm-hmm.

There's a lot about Kubernetes and stuff that is too hard. Someone should invent some sort of Windows interface for that. Right.

Or, or get all these engineers together to work On, get ejected AI And they all have a common problem. Sell More cloud s This, this stuff is too hard. It's too complex.

By the time I stand up the whole stack, you know, I gotta be like, you know, I need five geniuses to do that. I mean, Well that brings me to another announcement today. What's it called?

The Golden Coob. The golden uh, Berau or something. Golden Ber Knot.

Did you read about this one yet? Well, here's the deal times on this one. If you go take 16 or 14 or 16 certifications.

13 from the CNCF and one from the Linox Foundation. Okay. 14 certifications.

They give you like a won or golden ticket. I don't know what you do at your Wonka, but a golden ticket. And you get 50% off to CubeCon for Life And a swag.

You get swag, you can give, Uh, you get discounts on renewing your certificates or you can give those discounts to your colleagues. Yes. Or you can, and then I think they have, uh, oh, you get like 50% off for one of those, uh, you know, days they have around the world where they, You also, right.

You get discount certificates for more education too. Probably no better than a t-shirt. You get a golden eyes image of yourself on their website.

Oh wow. So now thinking that through, like how much, I'm Not kidding. It's true.

I'm not kidding. It's true. And you can't Make this up Get, and you get unique clothing that only you can wear, right?

Apparently. Yeah. I was, I you're gonna say like, and you can charge your own tariffs on, you know, source projects or something.

So, So here's the thing. I Charge only I could charge tariffs. Uh, here's the thing.

I was trying to figure out like how much time and every would it take to get those certificates Golden. The Golden. Because a lot of them actually, you have to show and demonstrate hands-on capability, right?

Like half of them are kinda like, you know, log in here and take the test and, you know, Sounds like CCNA at, at Cisco, right? But I mean, it sounds like one of those kind of, But but many of them are like, you actually have to show hands on expertise so you can't fake it kind of thing. I I imagine there can't be more than a handful of people.

Well, right. I think right now they said there's 14 of these people who were, um, these new class of Golden Cosmic not cosmonaut. That's Russian Kuber Kuber.

Is there a secret ceremony where you can make it into the, I know higher echelons of This now. There are other Kubernetes, these are golden Kubernetes. Kubernetes they launched last year is like, I think you have to get five certificates for that one.

So a gold What do you like a, an iron Bernet at That point? Or I haven't said anything about, you know, whether these are Brass, a brass kuber. It was like super sugar kuber knots or something new cereal from No, no, this is real.

This is real. Okay. Alright.

Hey, congratulations to those 14 Golden Kuber knots. And I cannot wait to see how many more people join the ranks of the Golden Kuber knot between now and let's say Atlanta, I think. And I think there's an actually like 1100 Kuber knots already.

So, But not golden. Not golden. There's only like 14 of us.

There you go. You heard it here first. The things you learn on Textron Gang.

We're gonna take a break. We have a third block if you wanna stick around. We're here at cucu.

com is the leading resource for news analysis and education on challenges facing the cybersecurity industry. com covers all aspects of cybersecurity, including data security, DevSecOps, cloud security, application security, network security, security threats, and more. com has the largest selection of security content featuring breaking news, blog posts, podcasts, and more.

com to learn more. com. Home of Security Bloggers network.

All right folks, we're back for our third segment. In no trip to Cloud Native Con is complete without the discussion of Observability Day, which again took place this year. And coinciding with that, Grafana Labs has their annual report out and it shows that well, 75% of folks in an open source community are using Prometheus and Open Telemetry, which may not come as a surprise given the open source community here.

But one of the things that came clear in the Grafana survey is that people were really cost sensitive about observability. And the issue was that I don't know if they love Prometheus exactly, or if they just feel like, you know, it's the other platforms are too expensive, I gotta hire people to set 'em up. I gotta curate all the data.

And, you know, Grafana is making a case for more of a stimulus approach that doesn't require you to curate data, but you've been tracking this space. What's going on from your perspective? Well, so, so observability is a data management problem, right?

The massive amount of data that's being collected into one observability platform or whatever the product is. But that also means transformation of data, right? Because it's all not all coming in in native format out of Open Telemetry, it's going through other agents and tools that you're combining all this stuff.

That's a lot of work. You know, developers don't like messing around with just overhead of transforming data. So I can see why people would stick with just open inflammatory and Prometheus and then say, now, now you might ask why Raffa is interested is putting a dashboard on top of that, but I, I'm not surprised actually to hear that.

That said, there are many others. I mean, it's a big market for, for observability, so it's not like everybody's doing this and only a few are running a third party commercial product. So is it really the exception or the rule?

So, oh, no, no. I was doing my kung fu Oh, grasshopper Ian, whatever his, okay. Um, I have, I have a differing opinion on that.

Okay. Number one, I just wanna point out that though, yesterday there was a observability day. It was not the biggest, uh, event of the, uh, satellite events yesterday, but from engineering was, which I think is quite a Mark.

com. But that being said, the issue with observability is yes, data storage is expensive. And where everybody's always trying to slice that somehow the real reason why they're all using, uh, open Telemetry or Prometheus is because none of the commercial providers have distinguished themselves from the other commercial providers to make a good case.

Why you should pay those exorbitant fees when they're all Prometheus and open telemetry under the covers anyway, if there was, if there was a really great observability platform, you wouldn't see 80% plus of the market using these tools. But I think a lot of engineers and teams look at it and say, if I just use the open source tools, no matter how clunky or harder to use they are, I'm getting 80, 85% of the functionality I get from these commercial vendors that are an arm and a leg. And for that kinda money I'll suffer a little bit.

Don't you think we're in an untenable situation? Because if Tenable's not an untenable has an observable, Yeah, well, yes. So It's not un Not unintended, but untenable a situation.

I mean, how many observability platforms products are there? I mean, it takes two hands to at least to count. Yeah.

That, I mean, but that gives that another indication that none of 'em are that much better. Well, Why, but why hasn't that sorted out yet? You know, why aren't we seeing, of course.

Yeah. Splunk got bought by as a security platform for Cisco, but you don't see observability companies getting bought left and right or collapsing, folding in. Are we headed to that?

Is that where it's headed? Well, According to Grafana, there are actually 101 technologies that are technically come under the heading of observability. Oh my gosh.

Uh, how they counted that, I don't know. All right. But Is that like dalmatians?

I, I guess maybe. But the issue in my mind becomes, um, it's the tools. I think people are using open source tools.

'cause I don't have to sign a check for somebody. I don't gotta go get somebody to approve that the developer downloaded Prometheus and started using it. And that became kind of a quasi standard.

But the problem is, whether it's the commercial ones or the open source ones, by the time I get this thing rolling, I need data engineers. I need developers to go play with Grafana dashboards and set all that stuff up. And then I gotta maintain all that on top of that.

This is like a DevOps project that lives forever just for this one little task. And I gotta ask myself, do we really have the right answer here for observability or have we kinda like been throwing crap at it all this time and not really solving the problem? Open source isn't this magic nirvana of now I don't need all these people.

You need actually a lot more people They do as opposed to outsourcing it. So, But Mitchell people have always been shortsighted. They look at the price tag of the software, not the total cost of ownership.

How can I get, how can I bring this in the door? Yeah, exactly. Because Well, employees are a sunk cost man.

We're a giving, I get it. I get it. I'm just trying to think picture to myself.

You asking yourself self, why are we doing this? I'm not sure I ever wake up and have that conversation. Right?

I think most people go to work and just keep doing the same thing over and over again because that's what they were hired to do and inherited. I don't think they, it would be, I think we, there's a lot of things in it that we need to take a giant step back and say, Hey, are we doing this? Because that's the way we always did it.

Because that's the answer. It's probably wrong, But I, I think nevertheless, it is what it is. I a lot of people do that now, Don't you think this is really a plot by the cloud providers to drive more business into Google Cloud?

But if it is, it's un-American and I'm not gonna be part of it, but then they, you know, They just build more agents and It's a whole universe inside your thumbnail. Me. Oh wow.

Oh wow. So you're saying this is a cloud storage conspiracy? Yes.

Yes. Well, look, I, I do think storage is an issue. It has been an issue continues to be an issue.

I think there are some companies with some interesting, like read it on the fly and store what we think you need, but who, who trusts them. And the other issue is the truth of the matter is you don't need to store all this data. There's only a subset of, of the telemetry data that you need to keep for any kinda long period.

And if you do need to keep it for a long period, to your point, you throw it up on an S3 bucket somewhere. But the amount of data that you need active for observability is relatively small. But nobody has the time to go and figure out, you know, exactly what logs, what traces and what metrics do you, This sounds like an AI job, an AI problem solve.

It could be, it could be another thing for an AI agent Or cloud or AI agents. We're back full circle. I hired before AI agents is the answer.

What was the question? Doesn't make a difference. Stay tuned for tomorrow.

Will the squirrel buy it? Will Natasha and Boris know? You know, I don't know what we're gonna talk about tomorrow, but it probably won't be AI agents and selling more cloud.

We hope you've enjoyed this. Uh, where are we doing tech Strong yank here at Cube Con. We've got a, a not a lot more coming your way.

We, uh, stay tuned. We'll be here for two more days. But for now, on behalf of selling more cloud services and an AI agent coming to you, I'm Alan Shimmel and we're out.