Platform Engineering Insights – David Sandilands, Puppet

This is techstrong tv. It a great pleasure of being joined by David Sandlin, who is Principal Solution Architect with Puppet. Welcome, David.

Hey, Mitch. Good to be here. Good to have you here.

Say, you know, I, I, I've been a long time follower of the, uh, DevOps state of DevOps report, uh, coming from Puppet and, you know, user of Puppet, et cetera, o over the years. Um, and it's great to have you on. You released it earlier this year, but I know that this year you focused more directly on kind of some of the merging things coming out of DevOps since we've been doing DevOps for a while.

And that, uh, was what you spoke at, uh, spoke about at Platform Con around platform engineering. So maybe if anybody, anybody hasn't read the, the State of DevOps report, tell us a little bit about that then let's jump into kind of what you learned about, uh, platform engineering. Sure.

So, I mean, h historically, so Puppet's been doing the state of DevOps report for, uh, over 10 years. Um, historically it's been a much wider report focusing on all areas of DevOps, you know? Right.

You know, it kinda started, you know, just a few years after kinda DevOps was a thing, and it was looking to, uh, find out what organizations were doing and what they meant by DevOps. Since, um, DevOps can be a very broad term, depending on organizations. Uh, and it looked to measure the maturity of different organizations, um, and, and to try and kind of bring out those patterns of what mature DevOps organizations were doing.

Uh, what we found over the last kind of, uh, three, four years of reports was that, um, although organization there was less, um, low level maturity organizations, uh, very few were actually kinda making it, you know, the, the, the growth of, um, highly evolved organizations was, was not going much further. So there was just the majority were kind of stuck in the middle. Mm-hmm.

Um, so as a result, we, we decided to focus down into those results and look at what, um, the standout features of highly evolved organizations were. Uh, and platform engineering had become, become a hot topic. Uh, and it was also seen in the 2022 state of DevOps report that, uh, platform engineering was one of those common features that, um, highly evolved organizations had more, um, the middle or low level organizations.

So the 2023, uh, report focused, it was a smaller report than the previous, uh, previous editions. And it focused just down on platform engineering to find out, um, you know, how long organizations have been doing platform engineering, uh, what their kind of focusing it was. Um, and you know, what, how they were finding the successes in terms of, uh, speed of delivery, um, and, you know, features like worthy using product managers and so on.

Mm-hmm. Interesting. Yeah.

Very good. So, so talk a little bit more about that. I think, by the way, it's very insightful to, to focus on that and uh, kind of dig a little deeper into merging Yeah.

Strong trends like that. What were some of the highlights that stood out to you that you were like, Hmm, I expected that, that's good, or maybe I didn't expect that. That's interesting.

I think one of the most interesting things was for, for something that I think some people are maybe a little bit tired of, cuz you know, platform engineering has become the latest buzzword that people like to throw about was that, um, there wasn't much cynicism in terms of the, the, the, you know, the, I think 94% of respondents said they found it was a positive thing for the organization. Um, and a similar, um, majority found that, um, that the aim was improving the speed of delivery for the organization, which was very interesting because the majority of organizations had had, um, their platform for less than five years. And when we were looking at kind of organizations who only have done DevOps for five years in the past, we weren't seeing that kind of impact.

Mm-hmm. You know, over time DevOps has had, you know, ha has caused a great improvement in delivery speed, but there was certainly a lot of issues for organizations starting out and a lot of organizations did not feel impact. Whereas platform engineering seems to be having an immediate impact.

Uh, and it also, um, shows good positive growth in terms of organizations are two to three years. Uh, you know, we see a big improvement as they go kind of three to five and above in terms of, um, you know, uh, speed of delivery and confidence that, uh, the platform will keep delivering. Mm-hmm.

Um, the, the, one of the other kind of interesting big differences for, um, platforms that have a higher, you know, higher delivery and higher confidence within the organization is those that, uh, have a product manager. Um, and I think that's, you know, one of the most important things we were taking out the report is that the most successful platforms do, you know, do it properly and have a focus on having a product manager who's evangelizing, um, who is, you know, focusing on the future of that platform. Uh, cuz my general concern would be since we got a lot of young platforms, uh, you know, you've, you've kind of got that initial boost when there's a, a digital transformation program or something like that, and then it's, you know, how does it kind of survive after that kind of five, you know, that initial positive period where typically a program sponsored.

Interesting you mentioned that cuz I just, uh, finished talking just recently with, uh, uh, Lambrose Caris who with Wise, who is a product manager for platform engineering at their organization. So that very topic, which is sort of, you know, it, it's good that we're not just doing platform engineering cuz everybody else is doing it. com, that we should self self plug there.

We should, we should be doing this, so let's go do it. Um, I'm, I'm curious, why do you think that we are kind of more hitting the ground running with platform engineering where we kind of have a lot more field to plow to really get ready to be productive at doing something much maybe as broad or bigger as DevOps? Yeah, I mean, I think the thing is DevOps was so broad and it was such a big change.

You know, my, my own background coming, you know, working in a bank, uh, historically before Puppet, you know, the, the methods of working before those kind of organizations went to DevOps was, was very manual. And I think the, the, the change was far greater. Um, and I think once a lot of DevOps work had been done, you kind of, you, you kind of had kind of groups of teams who had varying expertise and were having to kind of provide a lot of things for themselves in different little divisions within departments.

Mm-hmm. Uh, whereas I think, you know, immediately the, the cause platform engineering was more prescriptive, it was more focused on doing an exact thing. Uh, the teams were immediately able to kind of gather around that and those kind of things were, DevOps sort of stalled for a lot of organization.

So platform engineering was immediately able to focus on and pick up. So, you know, the most immediate things in terms of like infrastructure provisioning, um, I, I think that was, you know, that's probably been the biggest gain for a lot of organization mm-hmm. Is just having that kind of, um, almost like an, you know, particularly for internal organizations for their, you know, their own kind of, um, AWS or Azure and their private cloud where, you know, there's been a clear model set out by public cloud providers and then a lot of organizations internally would then able to kinda say, you know, that that's what our platform's competing with.

And that's what people expect now is to be able to self service, to be able to have these clean interfaces and um, you know, just be able to order infrastructure. Mm-hmm. You know, I think maybe this is just to guess my, my part.

Maybe we aren't in the, uh, disillusionment of hype around platform engineering cuz not everybody's co-opted it yet. Right. That typically is what, what happens in our industry, right?

Yeah. Yeah. And that's, that's what I think will be interesting for future reports is that, you know, the, you know, as I said, a lot of organizations in that one to five year period, there's typically gonna be, you know, a chief of technology or an executive sponsoring a digital transformation and a platform's gonna be key to that.

And I think as those platforms become older, I think that's where the challenge is gonna become. You know, I, I have seen, uh, organizations where, you know, the product managers been cut or, you know, they've decided to kind of slim down the platform team and, you know, that'll be a struggle for them to continue to evolve the platform. Um, you know, and that's when people kinda say, you know, platform engineering is nothing new.

You know, there, there is certainly a, a degree of truth to that. Um, but I think it's, you know, it's not had this kind of, all the components that make up platform engineering are all, are all old, but I think bringing them together and following the discipline is what new, what's new. And I think that's what we'll see going forward is which organizations kind of, um, I dunno, go from good to Great in terms of that things of continuing to focus on it and to continue to fund it and get the benefits from funding it.

Well, there's a difference between we do all those things versus let's bring it together, let's focus on it, let's make it a discipline, you know, maybe get more visibility, which oftentimes means accountability to it too, right. So there is, there's a difference between just we've been doing this versus it's a thing. Yeah.

And I, I, I think it's a thing as well as, you know, genuinely committing and training people into those roles. You know, we mentioned product managers, you know, it's, it's that classic thing. I've been in teams where, you know, the, one of the technical staff has just chosen to be the line manager because, you know, they need a line manager, so you end up somebody with two jobs and that, you know, that'd be our concern for platform engineers and, and product managers is, you know, it shouldn't just be something, you know, you've, you know, your consultancy's told you should be doing platform engineering.

So you just kind of pick a bunch of people and say, Hey, you're now platform engineers and you're the product manager. If, if it's not done as a serious thing and it's not invested in and people aren't trained and you know, you don't genuinely adopt the culture, you're not just gonna get the benefits Right. On team that they called that.

Oh, you were holding the book last, you are the platform engineering expert. Yeah, exactly. For whatever topic.

Right? Yeah. And I'm curious, David, so, um, from a puppet standpoint, how does platform engineering and that kind of thinking and discipline, how is that informed or maybe shaped your kind of what Puppet does and how you work with your customers or maybe where, where things are heading from that standpoint?

I'm curious your perspective. Yeah, I I think it's very, you know, I was speaking to our support teams recently and we were talking about, you know, cases tend to be a lot more complex now. Um, because what, what we're, what we're seeing is that, um, you know, puppet is now being far more integrated into other tools.

Um, you know, the, we see a lot more, um, a lot more focus on using puppet's, APIs well into things like Orchestrator to run tasks. Mm-hmm. Um, into things like ServiceNow where we have kind of, uh, built in integrations.

Um, and, you know, there's a big focus on, you know, kinda metrics monitoring observability where organizations really need to understand the capacity of these platforms. Cause I think if you went back to when I first started using Puppet, you know, 10, 10, 12 years ago, something like that, um, you know, you had a relatively static thing, which people were kind of, you know, by hand were coming in and ordering servers and, you know, you got something at the end, and if that thing took a hit, it was a pain, but you know, you got coffee and you got over it. Uh, whereas now, you know, you've got this living platform that, um, has so much more going through it and is, you know, is dependent on so many different self-service systems, you know, uh, bu you know, not just building infrastructure, but various tasks around that in terms of, you know, organizations who are allowing people to self-service, patching and self-service, uh, deployment of applications better.

Uh, so I think the whole thing has just become a lot more complex and a lot higher volume, um, with a, you know, with far greater dependency in these platforms than we ever had before. Yeah. The number of use cases is, is, you know, multiplied so much, which means the value of what you're getting from it is also, you know, greater.

I, I, I think it's always a funny thing when we talk about automation. We often talk about man hours saved, but I, I often think it's almost kinda false idea because you, you know, it's not like you're gonna say, well, there's all these things we don't do anymore. It's actually kind of, it just moves it along to now we can do all these extra things.

Um, and, you know, we've got a far greater capability of, um, you know, being able to rebuild labs quickly or being able to, you know, be patching as a service, which once used to be kind of a, a very hands-on job for people. Um, you know, it, it, it just kind of keeps building the complexity, which means people have to stay on top of it. Uh, you know, and coming back to that point of, um, you know, investing in the teams correctly, that would, you know, again, be my concern is, you know, when you've got successful platforms, you're gonna create that growth.

You're gonna create that hype. And as people discover them and realize what they can do by self-serving APIs, they're gonna find crazy complex things they can suddenly do. Um, you know, that they can just kind of throw servers up and bring them down as quick as they like.

Um, and, you know, a lot of puppets, um, components being kinda stretched by that at points. You know, we, we, you know, we have customers who are now kind of doing, you know, in instead of doing containers at points, they're just still using VMs, but you know, they have thousands churning around, you know, which was never a number we initially looked at, but it's something we're having to adapt to and, and to get used to. Well, very good.

Thank you much, David. com site. Is that correct or?

Yeah, if, if, uh, mean certainly if you Google, uh, pop it by Power Force, uh, state of state of DevOps, uh, platform engineering edition 2023, you, you'll find the report, uh, it's free to download, uh, and has the full summary. Fantastic. David, David Sandlin, who is, uh, principal Solution Architect with Puppet slash Perforce.

Thank you very much for joining us. Hope you come back soon. Perfect.

Thanks having me on, Mitch. You Bet.