Over a meal at Sune in London, RedMonk’s James Governor and Alois Reitbauer, Chief Technology Strategist at Dynatrace, use the restaurant menu as a lens to discuss one of the thorniest questions in enterprise software: when to build and when to buy? Oysters, charred flatbreads, and anchovies become stand-ins for open source components—widely available, broadly standardized, but rarely as good when you try to prep them solo. They chat about why packaging and end-user experience have become the decisive battleground in observability. James shares RedMonk’s long-held view that “the best packager wins and wins big,” and the two discuss how OpenTelemetry’s rise as an industry standard has shifted competition away from data collection and toward opinionated experiences on top of it. Along the way they touch on AI as a wrapper, the fast-food-versus-fine-dining spectrum of customer needs, and why the job of an enterprise is to use an observability stack, not manage one.

This RedMonk video is sponsored by Dynatrace.

Transcript

James Governor (00:00)
I mean, I don’t know that everyone’s going to want to be thinking about software as anchovies, but at least for me, I know they are really

James, Alois, we’ve had a good workout. We’ve earned some food. So the task at hand is to understand a menu and to get excited about it and mostly to realize that though these are just ingredients, they are things that if we made it, probably would not be as good.

Alois Reitbauer (00:42)
Plus just by knowing about the ingredients, we do not necessarily know what we get exactly. Because it’s not all of them. It’s just the main ingredients, not the entire dish.

James Governor (00:51)
is a bit

of a packaging exercise. So you don’t know what it’s actually gonna look like. We know what oysters are gonna look like. But here’s the thing. So oysters, I get oysters anywhere, but getting it right is still not always easy. I find one of the things with oysters is, and this is like the enterprise saying, I’ll do it myself. I’ll get the oysters and I’ll prepare the oysters, right? Which sounds great, but then you get the oyster and it has that little bit of shell in it. Doing it yourself is not always a good idea.

Alois Reitbauer (01:21)
Plus

oysters, mean oysters can be a lot of different types of… So, likely it’s gonna be…

James Governor (01:23)
Okay, there you go.

These

are good British golden oysters mate.

Alois Reitbauer (01:32)

That’s like sticking to the manual, just saying always as it could be.

James Governor (01:37)
And then a charred flatbread. Sure. Like, are you going to make your own flatbreads? Like you can buy the flour, you know, you can, mean, what is in a flatbread flour and water? And what is it? Yeast. I think there’s yeast. So you, in that case, you don’t really want to be making your own flatbreads, but well, maybe you do, but not everybody does. And I think this is one of the things you’ve got the open source components.

Alois Reitbauer (01:49)
it yeast? I think it was yeast. That sourdough is yeast.

James Governor (02:06)
things that you can work with, but whether you really want to be yourself, making them? Open question.

So here’s the thing we think about packaging. We talk a lot at RedMonk about packaging. You take open source software, but it’s how you package it and put it together in order to create a great experience for the end user. And we go so far as the best package wins and wins big. We think about OTel. Everybody’s going to offer OTel. It is an industry standard, but how you then manage and get value out of that, that’s something that’s a value to the customer, a value to the user.

Alois Reitbauer (02:38)
This is basically what the restaurant does. I mean the ingredients are not like open source but available to everybody like most of them you could buy. But what the restaurant is really about is a bracket packaging and experience.

James Governor (02:52)
That’s exactly right. It is that the end user experience, hopefully the developer experience, the operator experience, and in our case, the tasting experience. But I think that’s so important that whether or not we could do these things ourselves, could we really do them as well?

Alois Reitbauer (03:12)
And is it actually worth our time? think that gets also funny you look at what you’re building yourself and what are you taking pre-built or like…

James Governor (03:21)
Honestly, we see this again and again in open source and particularly in the observability world. So many enterprises build their own stack. They’re like, you we’re going to do it with Grafana. We’re to do it with Elastic or something. And then suddenly they find that they’ve got five people managing the stack. Then they’ve got 10 people managing the stack. Then they’ve got 15 people managing it. Suddenly you’ve got 25 people managing your observability stack because you put it together yourself. Sometimes I think there’s an argument that actually you should buy the package and,

You probably have some opinions about who that package should be from.

Alois Reitbauer (03:52)
Yeah, obviously you do have some opinions, but at the same time, I see like people doing both. Like you’re relying on open standards for where the data is coming from and how you’re collecting it. But then the packaging and the presentation is very slow to get opinionated. It’s like, what’s this menu here? If you agree on ingredients anchovies, olives, everything we have. These are good ingredients and we can agree on this is the best possible ingredient you can get out there. Like this open source, different libraries, you pick and choose the library that best fits your needs.

way you get really opinionated is how you put the dish together and I think that’s why you pick and choose who has

the right opinionated approach of what you want to do. And it might be different for different people, honestly. Some people want to just, okay, just show me the data. I’m totally fine with it. Other people say, okay, I want to have some more intelligence there. I have like run a very big estate, like some of our customers like running thousands of applications. And others say, well, actually I don’t want my people to do like the everyday work anymore at all. I want to have like the full automation, the AI capabilities and all of this in place. So depending on where you are on your journey,

would appreciate things differently like in a restaurant. There people who like fast food, there are people who like fine dining. And depending on where you are your journey and what your actual needs are right now, you might pick and choose differently. And there’s no right and wrong. There’s just more like, what is it that you actually need in a

James Governor (05:17)
But one thing that we’ve seen again and again, and it gets back to this packaging question at RedMonk is when we talk about things, it’s like, well, we’re in a transition constantly towards.

more managed services. I people don’t want to be managing it all themselves. There’s a reason why people historically bought packaged software. And today I think there’s a reason why they buy managed services. And when I say managed services, I don’t mean in traditional sense. I mean, software as a service and so on. that value proposition, and obviously some organizations have good reason to manage it themselves on-prem. But the…

The packaging matters. We see that again, I think in AI, so important, so important. It’s a wrapper. That’s the user experience. That’s what we’re here for. So to your point, it’s like exchanging money for time. We’re exchanging money for the experience. I mean, the job to be done of an enterprise is not to be managing an observability stack. It’s to be using an observability stack.

Alois Reitbauer (06:19)
I think as an enterprise you really have to look at your economies of scale. you’re building internal infrastructure it’s usually not an economy of scale. You’re building it, you’re running it. It’s one by one. Like for us if we build something we run it for a couple thousand of customers. So our economies of scale are much better so we can afford to obviously go much deeper there. And I think it’s all out of service. It’s not that we build everything ourselves.

you only build what differentiates your offering and the rest you don’t do and for most companies observability is not differentiating your offering the database you’re using is not differentiating your offering honestly not even like the networking stack or maybe kubernetes distribution is actually differentiating your offering.

James Governor (07:01)
I mean, you’re

Who really wants to build their own database? Or even to like, what do want to do? You want to like build all of this routing infrastructure for your Postgres? Thank you. Thanks so much. wait, I should have, I know be prepared. Sometimes you have to do some of the work.

Alois Reitbauer (07:18)
being prepared.

James Governor (07:27)
So that was an amazing meal at Sune on Broadway Market. Definitely visit if you’re in the neighbourhood. So great conversation about observability, open source, packaging and AI sponsored by Dynatrace at Sune in London.