Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
What is Platform Engineering with Luca Galante
1.1k Plays1 year ago
In this episode of Kubernetes Bytes, Ryan and Bhavin sit down with Platform engineering expert Luca Galante, head of product at Humanitec to discuss all things Platform engineering, what is it, and how it is different from DevOps. They also discuss when it makes sense to have a platform engineering team and how you can get started on your journey by building internal development platforms.
Show Notes:
- What is platform engineering?
- What is an internal developer platform?
- What is Dynamic Configuration Management?
- Platform Engineering community
- PlatformCon 2023
- Luca’s LinkedIn and Twitter
Cloud Native News:
- InfluxData Series E round - https://techcrunch.com/2023/02/13/influxdata-lands-51m-to-grow-its-time-series-database-offerings
- Cloud Native Security Con - https://youtube.com/playlist?list=PLj6h78yzYM2NQ-Zi_k5qVmZyxSmLBzM6V
- Adrian's Platform engineering blog - https://adrianco.medium.com/platform-engineering-teams-done-right-b3b3d4a8ad23
- Amazon EKS Anywhere on SNOW - https://aws.amazon.com/blogs/containers/new-amazon-eks-anywhere-on-snow
- Kelsey Hightower affirms databases on k8s - https://twitter.com/kelseyhightower/status/1624081136073994240
- NetApp discontinues astra data store https://blocksandfiles.com/2023/01/25/netapp-astra-data-store/
- Edge computing trends 2023 - https://technologymagazine.com/articles/the-cutting-edge-edge-computing-trends-to-watch-in-2023
Recommended
Transcript
Introduction to Kubernetes Bites
00:00:03
Speaker
You are listening to Kubernetes Bites, a podcast bringing you the latest from the world of cloud native data management. My name is Ryan Walner and I'm joined by Bob and Shaw coming to you from Boston, Massachusetts.
Meet the Hosts: Broadcasting from Boston
00:00:14
Speaker
We'll be sharing our thoughts on recent cloud native news and talking to industry experts about their experiences and challenges managing the wealth of data in today's cloud native ecosystem.
00:00:31
Speaker
Good morning, good afternoon, and good evening wherever you are. We're coming to you from Boston, Massachusetts. Today is February 21st, 2023. I hope everyone is doing well and staying safe. Let's dive into
Weekend Reflections and Super Bowl Highlights
00:00:45
Speaker
it. Bhavan, I hope you're doing better this week. I am, I am. Thank you for asking. I'm doing way better. Yeah, I think it was just one of those weeks, last couple of weeks back. I'm back to normal. Good, good.
00:00:58
Speaker
But yeah, coming back from a long weekend, so that definitely helps, right? Like this was a long weekend. The day we can always helps. And even though I regret the four day work week, because you have to do the same amount of work. Now you have four days supposed to be how it works.
00:01:14
Speaker
But yeah, I'm glad the weekend was good. And now we're back doing this podcast. Yeah. Did you wind up watching the Super Bowl or anything? Oh yeah. Like we did have a few people over. We did some Super Bowl squares kind of thing, but like not the typical squares, but we did some other over under bets with just like 50 cents or something like that per bet. That was fun.
00:01:36
Speaker
I was bummed that the Eagles lost although it was a really good game like it was a high scoring game you were engaged the whole time and like you were in it but not happy with the results but again the Chiefs won they deserve it. Are you an Eagles fan? No not really I just for the Super Bowl it was like Eagles or Chiefs yeah it's Eagles but that's it yeah I'm going back to supporting the Patriots now so
00:01:59
Speaker
let's go below Brian, right? So let's see if we can turn the ship around. Nice. I think I remember saying I would like the Chiefs to win because they're sort of the quote unquote underdog, right? So I know the Phillies fans, the passion fans. But again, I'm not an either or fan either. But it was a good football game to watch. Funny story, I went to a buddies of mine to like,
00:02:22
Speaker
hang out with our kids and eat some food and stuff like that. And he informs me. He's like, he goes, you know, I switched to streaming, you know, a while back. So I don't, I don't even have TV. Like how are we going to watch the game? And I was like, what do you mean? You invite me over for Super Bowl and we can't watch Super Bowl. Luckily, luckily Fox had that free app, which worked most of the time. Um, it froze a few times, but you could watch the game for free, which is pretty good.
00:02:46
Speaker
That's a funny way to invite people to work. Yeah, let's come over and then we do Sunday dinners and stuff with these friends and it's like we were there for like to be with each other and the kids like you know, so it was sort of a Secondarily like oh the Super Bowl is on cuz we're all like we don't all love the ball. Gotcha
00:03:02
Speaker
Um, anyway, it was good. It was fun. And I did some squares too with a neighbor of mine. I rarely ever do and publicly admit to gambling, but I won 200 bucks. I won the third quarter. Um, yeah, I put in 40 one 200, you know, I'll call that my, my yearly gamble expenditure.
00:03:25
Speaker
If you are scrolling through any social media accounts in Massachusetts, you are bombarded with those draft kings and fan dual ads because now it's legal. Everybody's like, let's go. Yeah, I know. That was legalized like last year, right? I think late last year. I forget. Yeah, it was an ongoing process. I think right now in February, you can bet on sports and through casinos. Maybe in March, I think you can do the mobile app thing.
00:03:51
Speaker
But yeah, I don't know, it keeps changing. And again, since I don't gamble on sports, I don't really follow that much. Yeah, no, I hear you there. But I know a lot of people do. So power to them. I had a buddy win like a thousand bucks or something like that and put it in like five. So I was like, your ratios are much better than mine.
00:04:09
Speaker
At that point, you should just stop, right? Take the $1,000, that's it. You can tell that story forever without losing any more money. The thing he won was he picked, in order, the first three people to get touchdowns or something like that. Nice. The individual people. The odds were pretty hard.
00:04:27
Speaker
I was like, good. He's also like, he used to be a sports writer. So he's like, okay. Besides the point anyway, we have
InfluxData's Funding and Cloud Native SecurityCon
00:04:37
Speaker
some awesome news. We have an awesome guest. And before we introduce our guest, we shall get into our cloud native news of the week unless you had something else. Let's dive into it. Bye. Why don't you kick us off?
00:04:49
Speaker
Okay. So I have a few news items today. I think the first one being a funding round. One of our previous guests actually in flux data, they raised a series E funding and that it was a bit complicated round. So they picked up $51 million in as part of the equity round, but they also added $30 million in debt. So like the total money raised was 81 million in this round. But then I see 31, 30 million was just a debt that that was raised.
00:05:17
Speaker
I think this followed, they did a series D in 2019 around 60 million. So I think they want to grow that on-prem base. They also have a new database engine. Again, you can read through the details in the show notes. But I think this brings your total capital raise to 171 million. That's good amount of money.
00:05:34
Speaker
Yeah, this 81 million helps them like innovate faster and then obviously stay around for longer because I know some of the big names on their website do use influx data or influx DB for Running databases on Kubernetes even so yeah, absolutely
00:05:50
Speaker
The next couple of articles were just things that people can watch. So cloud native security con, that was a separate conference, separate event this year in Seattle. It wasn't a day zero event at KubeCon anymore. The sessions for that are public now on YouTube. So we'll have a link to the playlist, which lists all 88.
00:06:09
Speaker
sessions, including keynotes. I've been watching a few. I think I like I really like the one that Liz Rice did about EBPF and how she used the Death Star and Millennium Falcon as kind of the intruder. So and how you can have security policy. So that was a fun, fun session, fun 20 minutes, if you're looking for recommendations.
00:06:30
Speaker
But apart from that, the second, the third thing that I had was more around the topic for today, platform engineering.
Platform Engineering Insights and Innovations
00:06:39
Speaker
I know there has been a lot of activity around this topic. Again, we aim to break it down by talking with Luca about it, but Adrian Cockroft, right? Like I know people really trust what he says, given his experience. He has written an article about platform engineering and how you can do it right.
00:06:55
Speaker
And then he has listed like four different principles of how you can build a layered approach for platform engineering. And then at the end of the blog, he has an awesome map key, which is kind of a mind map for what are the different phases. So what can be treated as utility or commodity, what can be treated as products, or you can rent things out from vendors, and then what's custom to you and what are the benefits that your team can get.
00:07:24
Speaker
that mind map was awesome too. So those articles will be in the show notes as well. Yeah, absolutely. Yeah, if you haven't read what Adrian has written elsewhere about containers, microservices and stuff like that, also highly recommend everyone that comes from him. Deep dive into that. We'll put some links maybe in the show notes.
00:07:42
Speaker
Oh yeah, that's a good idea. And then the final article I had was talking about Amazon EKS and EKS Anywhere specifically. Now they added a new deployment model. So I think before this, they supported EKS Anywhere on VMs, on bare metals. Now you can run EKS Anywhere on AWS Snow instances. So if you wanted that edge deployment, you didn't want to buy any additional servers. You can order your Snow devices.
00:08:04
Speaker
from your AWS console, tell it that you want to run the AMI that supports EKS Anywhere, and then load the EKS CTL Anywhere CLI utility. Once you have those new devices, you can spin up your cluster, maybe run this in a completely disconnected scenarios as well. So that was an interesting blog, interesting features. For people looking to run Kubernetes at the edge, this might be an alternative.
00:08:28
Speaker
Oh, look, we both put an edge article in. I'm going to do that one first, then. It was my third one. But I think general call to action. We'd love more edge topics on the show. And I also had an edge article that talked about the edge computing, I should say. Edge computing trends to watch in 2023. And it's a generic article. Nothing crazy in there. But it calls out containers being key and the use of cloud native storage at the edge.
00:08:55
Speaker
Um, you know, since we cover a lot of those types of topics here, I'd love to hear, you know, is that the way you're thinking? Are you using it that way? You know, um, you know, maybe you have a use case already, uh, you know, we'd love to come talk to you and, um, kind of build on this. Cause I think it's starting to, you know, we're starting to hear a lot more of that type of use case with sort of the.
00:09:13
Speaker
the trends and repatriation as well that we've seen. Cool. One of the things that comes to my mind when we talk about Kubernetes on the Edge is the Chick-fil-A article. I don't know if people have seen this, but again, I'll find a link how they are running these Nook-based Kubernetes or K3S clusters at the individual Chick-fil-A locations and how Kubernetes is helping them modernize. Just makes Chick-fil-A even better. Yeah, like you need another reason.
00:09:38
Speaker
I've honestly never gone to a Chick-fil-A ever. Oh, my wife loves them. There may be some haters out there for me. I've eaten it. I have an uncle who brings it to me. And then we're like meet up in the middle. He's like, oh, you have so many Chick-fil-As out here. Anyway.
Kubernetes and Database Viability Discussion
00:09:58
Speaker
The next piece of news was, we didn't cover it last week, but I think it was out last week was NetApp discontinued Astra data store, that is.
00:10:08
Speaker
to be, you know, very specific. Astro Datasaur was sort of their abstraction, common sort of control plane, and it was out in beta, I believe. I don't think it made it out of beta. But basically with, they're claiming basically it's been replaced by what they're doing with Blue XP.
00:10:25
Speaker
Definitely something worth taking a read at and what they're up to with BlueXP and all their platforms. Lastly, just to add to Astra Datastore, the Astra control service and Astra control center, the data protection play that NetApp has is still around. Just the cloud needed storage or the software defined storage solution that they were working on.
00:10:49
Speaker
in data store might be like that was the thing that was discontinued, but you can still use their Astra Trident plugin to talk to your on tap system.
00:10:59
Speaker
Yeah, it's a portfolio and there were so many different terms. So I just wanted to. There's a lot of different terms and a lot of them have the word Astra in it. They also renamed their CSI plugin from NetApp Trident to Astra Trident. So that's still part of the Astra portfolio. Yes. Thank you for the clarification. Cause yeah, I feel like I know it very well now at this point, just following the space, but it can be quite confusing. I appreciate that.
00:11:22
Speaker
The last link I have here is a tweet, actually, from Kelsey Hightower, which many of you may know that name. If you don't, go take a look at what everything he's done for this community. Very, you know, great person. And he did tweet the other day, I forget what it is, the 10th or something like that, that basically there was a comment of, you know, someone reaching out to Kelsey saying, bust this myth for me, running any sort of database instance on Kubernetes is a bad idea.
00:11:52
Speaker
And so, you know, coming from the background that myself and Bhavan, obviously we do not think that way, but it was nice to hear someone like Kelsey in the community say, you can absolutely run databases on Kubernetes because it's fundamentally the same as running a database on a VM, in quotes. And the biggest challenge is understanding that running it on Kubernetes won't be like turning it into Cloud SQL.
00:12:17
Speaker
And I took that also in quotes, but I took that as basically running a database on Kubernetes isn't like a full blown database service from a major cloud. And I actually tend to agree. I think there's different use cases for those type of things. You might choose one or the other. I think we've probably covered a little bit of this in some of our previous episodes. Maybe this is a good episode to have really to break down when you kind of
00:12:44
Speaker
use one or the other or that kind of thing. If you have comments, go check out this thread, though, because there is a very active thread. I think there's a lot of good things to hear from all sides of the coin. I mean, there's there's folks out there that had a challenging experience with running databases on Kubernetes and there's people that have been like doing it for five years, you know, no problem. So I think it represents the community and also just the problem statement that, you know,
00:13:09
Speaker
the whole sort of communities and data community and running databases and doing things with data on communities sort of tackling and near and dear to our day jobs for sure. So I think we have spoken to and experienced this in even in our day jobs, right? Like at the DoK day panel that we did at KubeCon last year, even the questions were people who have been doing state
00:13:37
Speaker
tried it in like 2017 when the CSI spec wasn't as evolved and have the volume snapshot features and they struggled with it and then they're like nope I'm never going to go back and after like having a conversation that they were like okay maybe maybe it's time to like reevaluate things things have changed in the past five years as you said it's I do want to like get a coffee and then read through this thread
00:14:00
Speaker
I highly recommend doing it, as well as anyone out there listening to this. Go check out this thread. If you're interested in this space, obviously, it's a good conversation between a bunch of folks in the community. And get involved, right? I'd love to see your comments on there, as well as mine. I think I put one on there. So, we'll see me chiming in.
00:14:20
Speaker
Cool. I mean, that was the last of the news and we do have a really awesome guest. His name is Luca Galante. Some of you may know him from his work at HumanaTech around DevOps. He is a product at HumanaTech and we're going to talk everything sort of platform engineering.
00:14:40
Speaker
That's a term that we've heard a lot. We mentioned it in our news. Platform engineering is a term that really just exploded in 2022, especially on KubeCon's, some in the limelight, some not so good limelight. So we want to understand what is it, what's the goal around it, and dive into it.
00:14:59
Speaker
But we do have a fun segment, which I believe if you listen to the entire last episode, we asked our guests to this segment. We came out with this idea prior to this recording this guest, so we don't have it directly asked to our guest, Luca. But the whole segment is basically we ask chat GPT to come up with a question related to the episode that we can ask.
00:15:25
Speaker
are a guest. And if they have an answer, that's great. If they don't, we just asked chat GPT to answer its own question, which is always quite enlightening. So this week's question that it came up with is, if platform engineering was a theme park ride, what would it be and why? And I thought the answer here, before I give you chat GPT's answer, do you have any thoughts on this, Papa?
00:15:48
Speaker
No, I'm just excited because I already know the answer. I wanted to share that. I don't think I can top that answer. Great. Yeah. So it is a little lengthy. So bear with me. So it says if platform engineering was a theme park ride, it would be a roller coaster. And here's why. And it's, it's, it thinks for a little while.
00:16:05
Speaker
Like a roller coaster, platform engineering can be full of ups and downs, twists and turns. Just as a roller coaster takes you on a journey of unexpected thrills and excitement, platform engineering can be constantly changing landscape with challenges and opportunities emerging every day. Just first paragraph, already nailed it in my opinion. The ride would have multiple tracks symbolizing the different paths that a platform engineering can take such as infrastructure design, tooling and automation, data management and security.
00:16:33
Speaker
The roller coaster would have steep climbs and sudden drops representing the high-pressure situations and unexpected challenges of platform engineering that they often face. It would also have thrilling loops and corkscrews symbolizing the creativity and innovation required to design and maintain a successful platform. At times, the ride would speed through dark tunnels representing moments of uncertainty and ambiguity.
00:16:56
Speaker
and the platform engineers would have to navigate. And just as a roller coaster often requires teamwork between riders to keep everyone safe, I don't know how true that is, it kind of does it all for you, but platform engineering requires collaboration between teams to ensure platform functions smoothly and securely. Overall, the platform engineering roller coaster would be exciting, unpredictable, and rewarding ride, much like a field of platform engineering itself.
00:17:22
Speaker
I don't know about you, but, um, you know, to the listener, you'll, you'll hear what Luca and us have to say. And, um, you'll see and judge for yourself how right chat GPT is, or maybe you're already doing platform engineering or claim to be, and it feels right at home with what the chat GPT said. Cause there's a personally a lot of truth in that. You know, my takeaway was like for this chat GPT is trained on 2021 data. Like it doesn't have all the changes that happened in 2022 and still
00:17:52
Speaker
It's so close to home. Or maybe it's a sort of a not as good as our representation. I'm not sure. But it is quite entertaining to hear the answers from this model. And you don't know what chat GPT is. Google it. Go check it out. It's really fun and interesting time to be alive, I guess they should say. All right. That's enough for our intro. Let's get Luca on the show so we can chat with him about platform engineering.
Luca Galante on Platform Engineering Evolution
00:18:20
Speaker
All right, welcome to Kubernetes Bites, Luca. We're excited to have you on here and talk about all things platform engineering. But before we do that, why don't you give the audience a little bit about your background and who you are.
00:18:31
Speaker
Yeah, thank you for having me. Yeah, so I'm Luca. I run product at HumaniTech. We built platform orchestrator. It's a core configuration engine to your attorney developer platform. And I'm also one of the core contributors to the platform engineering community. I moderate the Slack over there. I host PlatformCon.
00:18:54
Speaker
Which is coming up again this year off the con 23 and now you're writing newsletter platform weekly that has about 10,000 subscribers. Oh Wow, okay. That's awesome. Look, I didn't know you had a newsletter. I need to subscribe and I haven't done one of Lucas emails yet. No, not me The only person in the industry I'm not one in 10,000 I guess
00:19:20
Speaker
So, Luka, let's start by talking about platform engineering, right? Like, what is it and why should I care about it?
00:19:26
Speaker
Yeah. So, um, platform engineering for me is the art, cause it's really more of an art than a science of designing, um, golden paths. Um, and so in that sense, bringing together all the different tech and tools that you have in an organization into that golden path. Um, and that golden path, what it does, it enables developer self-service for developers and reduces their cognitive load specifically on sort of like the individual contributor.
00:19:57
Speaker
These golden pants are most often referred to within the community as an internal developer platform or IDP for short, which is the product that the platform engineering organization, the internal platform team, however you want to call it, ships to the rest of the engineering organization. Okay. You said platform engineering team is building this as a product, right?
00:20:26
Speaker
Where have we seen this trend coming from? What changed in the past couple of years? Why is this now a big thing? I think to answer that question, we need to take a step back and look at DevOps for a second.
00:20:44
Speaker
So when DevOps came up more than 10 years ago at this point, the world was a very different place. We were mostly developing monoliths on maybe bare metal. The infrastructure was considerably less complex. There was no Kubernetes, no IAC, no GitOps.
00:21:05
Speaker
insert your cloud native tooling technology. But the initial idea behind that is a great idea, right? Is remove the barriers between developers and operations, facilitate collaboration, you build it, you run it.
00:21:24
Speaker
That's all great. The problem is because we have this oldest converging trends, Kubernetes, containerization, cloud native, Terraform, and whatever, and we have this 10,000 tools CNCF landscape that keeps growing.
00:21:43
Speaker
Reality of the day-to-day of DevOps in a lot of organizations is actually not that great. In a lot of organizations, you actually have developers that are overwhelmed by these quite complex tool chains. And so something that used to be very simple back in the days when DevOps started, which is, hey, I just want to deploy a small change to an application and a preview environment and test something.
00:22:09
Speaker
That now, just to do that, it can take weeks in some enterprise setups because just to get permission to get that environment spun up and that database provisioned and yada yada, it can take days, weeks. It can be a very painful process. Or even if it doesn't have to go through all these gates,
00:22:31
Speaker
It's still a very overwhelming experience for a lot of developers. That's where we talk a lot in the community about cognitive overload because now they need to touch three, four different tools, three, four different scripts. They don't really understand this Helm chart or this Terraform module over there.
00:22:52
Speaker
I don't want to screw it up. What am I doing? All right, let me just reach out to somebody to help out. And so that somebody is an operations person on the other side of this like fence that or, you know, a DevOps team that is
00:23:07
Speaker
basically trying to put off fires and survive on a daily basis and fighting what we call ticket ops. In a lot of cases, truly become some sort of glorified help desk that is really stressed out.
00:23:23
Speaker
There is very low technical empathy sort of like that goes both ways. There's a lot of frustration, a lot of waiting time. And it's not just, it's just not that great. And so why does that matter? Because, you know, when this whole DevOps hype train started, and again, I think it has just some really solid foundations, but, you know, a lot of
00:23:50
Speaker
leading engineer organizations in the space or like all over the world, really. So you're talking about the Googles, the Spotify's, the Airbnb's and stuff. They immediately realized, hey, this is great in theory, but there is no way I can add hundreds or thousands of engineers every month.
00:24:10
Speaker
and expect them to all be familiar and understand my increasingly complex cloud native setup. There's just no way. And so they figured, all right, we need to build some sort of platform layer between the operation side of things and the developers so that developers can self-serve what they need to run their apps and services independently.
00:24:33
Speaker
And operations can focus on solving hopefully more interesting problems than setting up your end PostgreSQL or spin up your environment. And so that's kind of what we've seen. And then from there, and so this leading tech companies and organizations, they've been building platforms for
00:24:57
Speaker
almost a decade already at this point. From there, we've seen it slowly trickling down to smaller organizations and less advanced organizations. I think what the platform engineering community really has tried to do over the last two or three years is more than anything,
00:25:22
Speaker
help put a framework around all of these things, right? It's not like we invented anything. It was literally people were already building platforms. There was just no shared understanding of what it is that we're doing. There was no shared best practices, even just terminology. And so that I think was really just the value of the community was provide a safe space for people to share those worst stories and best practices. But I think more importantly, just give
00:25:48
Speaker
people a way how to think and how to talk about these problems. And I think that's why it really hit a nerve with this, you know, more senior platform engineer contributors that have been doing this for years. They just didn't know what it was called, right? So it was this big relief moment of like, Oh,
00:26:06
Speaker
That's what I've been doing. That's why I think it hit a nerve. If you had to sum it up, I think the key difference here that sets platform engineering apart from your SRE or DevOps is really this attention to product. One of the key principles that we advocate for in the community is the idea of platform as a product.
00:26:35
Speaker
And so the idea that, hey, you're not just focused on reliability, maintenance, and scalability, which is all great, which is why, by the way, a platform team does not replace your SRE team. It's just a separate thing. It's a separate effort. But it requires a completely different mindset and set of skills to really ship that product.
00:26:57
Speaker
Yeah, it's a really good point, right? And this is, I think, a topic that comes up a lot when talking about platform engineering, just because of the trend that we've seen in 2022. Moving into 2023 is like, how does it actually differ from DevOps? And you mentioned a lot of these enterprises still have ticket ops and those kind of things. I call it near DevOps, right? Because it's a lot of the time we strive for this golden star of
00:27:23
Speaker
being able to release quickly and all these things, but when you're operating at that scale, you can't quite get there. And also just the growth of all the tools and everything, it's almost, you know, a platform engineering, I feel like is an evolution of DevOps as well as sort of a new layer, right? So I guess this, you know, you mentioned a little bit about, you know, how does it differ from DevOps? I actually remember
00:27:48
Speaker
being part of a DevOps engineering team in 2014 at a company, and we were effectively building a platform. So you could argue that we were also a platform engineering team building a platform API, obfuscating a lot of the pipelines and everything that our development teams would do. So I'd like to dig into that point a little bit. Are there other characteristics or traits that make it different, or do you see it more of an evolution of DevOps in SRE?
00:28:17
Speaker
Yeah, so it's an evolution in the sense that is, I think it, you know, by building a platform, you actually enable true, you build it, you run it, which is the initial core tenant of doubles. That's really the key thing, right? Now, the how you successfully build and ship a platform, because there's a lot of people building platforms that don't go anywhere.
00:28:38
Speaker
Um, is the, um, you know, I think, I think is one, this, this product mindset, which means first of all, you need to build a product team. So, you know, in a product team, you have a product manager, you might have, you know, front ends, back ends, QAs, and so on, right? Like a regular product team. That is not the case in a lot of, you know, DevOps teams traditionally, right? And that's where we see, you know, it's important that you do that shift.
00:29:07
Speaker
You know really from a team composition perspective and skill sets perspective and then have this mindset, right? And the mindset is I think the most important thing because you really need to you know, put yourself in The position like hey, I'm building and shipping a product
00:29:25
Speaker
The same way as if I am right now as a startup, which means you also need to think about, okay, what is the problem that I'm solving? Am I building a tight enough feedback loop between my market, my internal developers' customers, and my product team?
00:29:46
Speaker
to make sure that we're listening to them, to make sure that we're iterating fast enough, and so on and so forth. That I think is super important. Also, just think about what is the distribution built in. This is something that might sound strange at first, but I think
00:30:08
Speaker
I was talking to Erickson, who's responsible for building the platform at Salesforce a few years ago. A very large organization is you actually have this interesting situation where you have multiple platform teams competing with each other, which is obviously what you want to see.
00:30:28
Speaker
But it also tells you, again, you really need to treat it as a product because you need to think about your go-to-market, essentially. And what is the team that I pallet this first? What is the dev team that I pallet the platform first with? And then where do I go from there? And how do I think about all the different golden paths? And you mentioned obfuscating. So how do I think about the different levels of obstruction that I build as I onboard more and more developers? And we can talk about that a little bit more.
00:30:58
Speaker
gets to is, I think the second most important skill, which is communication. I think if you were to plot from sys admin to develops engineer to platform engineer, you'd see an increase in what's the communication skill needed to succeed at your job.
00:31:20
Speaker
You know, back in 30 years ago when you were just like in your basement figuring out how to install servers and networking, you know, and they were just like throwing software in the fence and obviously that was part of the problem. There wasn't much communication skills needed anyways, right? And as we have, you know, done more and more, you just need this technical empathy, really, I think is a key term here to just listen to what orders in the org meet and how can you help them, right? And really put yourself in this position of like, I'm serving them.
00:31:50
Speaker
And so that's a key part of the skill set, I think, is communication with the developers, but it also goes towards other stakeholders. So executive buy-in obviously is really important. And I think a very tricky, also kind of like line of communication is with the pre-, with the already existing DevOps, SRE, sysadmin sort of teams, and how do you handle all of those sort of relationships.
00:32:19
Speaker
The concept of communication and organizational buy-in is imperative, I feel like. I've been a part of efforts where you could hire and build some of the best teams that have some of the greatest skills that know how to do DevOps and SRE, but if you're not built with the culture and communication,
00:32:40
Speaker
of kind of taking this on and having the buy-in from those stakeholders. Being a platform engineering team can be sort of a grind against the grain if that's not sort of in place for us. And I feel like that does happen quite often. Yeah. So my next question is around, again, your role at HumanaTech, where you're building these things, but
Starting and Scaling Platform Teams
00:33:05
Speaker
How do you work with customers or what do you recommend them to as a starting point? Like let's say I'm a small medium to large enterprise where I have multiple DevOps team at this point that are all responsible for building and shipping their own code. How do we go from that point in the modernization journey to building a platform team? I know when we started
00:33:25
Speaker
calling people devops admins, people were making fun of, oh, they are released as admins who now have a new job title and are getting paid more. Is this the same thing? Like I can call myself a platform engineer and get paid more because it's 2023? Yes.
00:33:41
Speaker
In fact, we did run, no, the answer is no, but we did run a survey recently that we published as the state of platform engineering. And we did ask in the community, and what came out is, for instance, in the US, I believe, platform engineers may serve like 20 to 30% more than their doubts counterpart. So it's already a halt. But to answer your question, I think
00:34:08
Speaker
I think it's important to distinguish between buying a plugin ready to go solution versus building a platform. If you're a small team, let's say you're 20 engineers and you are starting from scratch, completely greenfield, then probably building an internal developer platform and having a dedicated platform team is not the best idea.
00:34:38
Speaker
There's just too much overhead. It doesn't deliver a ton of value. You're not solving problems that you're really painfully feeling at this point. And so just go with the pass. It's great. There's a lot of stuff out there. Obviously, the predecessor is a Roku. There's a lot of new different passes for different use cases nowadays. So just go with that.
00:35:02
Speaker
Now, if we're talking about, okay, you're larger than that, and usually, you know, roughly rule of thumb, we say, in the community when you're about like 40, 50 engineers, more on a cloud native setup, that's when you start seeing things breaking, you know, those sort of like painful waiting times and frustration points that we were talking about earlier. That's when it makes sense to start thinking about, all right, here, here, we might need a platform. And so the first thing I think is,
00:35:33
Speaker
Understanding that you... This is really what we advocate for in the community is platform engineering is a toolbox. In the toolbox, you can have vendors like Humitec, you can have open source tools like Backstage and ArgoCD and whatever.
00:35:51
Speaker
But the key thing is you need to just pick the tools in your toolbox and then go and build that platform. And so the important thing here is to understand you as a platform team are the best positioned people in the world to understand what is the best golden pants and the best
00:36:14
Speaker
Um, you know, obstruction level for my teams internally. There's nobody else that can do that better than you, which is why we strongly advocate for, for this approach of kind of, Hey, he has an opinionated tool set to go and build opinionated workflows for your engineering organization. Because otherwise the flip side of that is the past, which is, Hey, here's an opinionated everything. And you can do very small tweaks, but that's, that's about it. Right. And that just doesn't scale.
00:36:42
Speaker
Yeah, this is a scale at a large at a large order. And so okay, once you understand that, where do you start? I think the first thing is, you need to have a very clear mission and role for the platform engineering team. Meaning, if you're just taking a DevOps theme and saying, hey, now you're building a platform, like that can work. But oftentimes, that doesn't work. Because the rest of the org doesn't really know what to expect, right?
00:37:10
Speaker
And again, your communication is really key. You really want to communicate, hey, here, our role is to build a platform to serve you as our customer. And this is how we're going to go about it, right? Versus like, yeah, we're also still kind of like answering your ticket op stuff, you know, but on the side, which, you know,
00:37:32
Speaker
It can be fine at the beginning, but the earlier you make, you clearly have a separation of concerns between not just devs and operations, but really within operations from more traditional DevOps tasks to platform engineering tasks.
00:37:53
Speaker
The earlier you just said, do yourself a favor, basically. And then the other thing is, okay, you want to treat your platform as a product, right? And we already talked about that. I think the other important thing is you want to focus on common problems, right? And this is where I think there is an interesting paradox almost.
00:38:15
Speaker
Because you might think, hey, I'm a brand new platform engineering team and I want to build a platform. What's the ideal scenario? The ideal scenario is Greenfield. I can start from scratch. I can dream up my ideal stack and off I go. I would actually argue the ideal scenario is the real scenario and the best majority of cases, which is you already have a complex Brownfield enterprise setup and you
00:38:43
Speaker
have, you know, different dev teams that have different preferences. There is this one that is shipping on Jenkins. This is one is using CircleCI plus Argo and so on and so forth. And so that already tells you a lot. And so you can already try to understand, all right, what is the minimum common denominator here to, to solve for, you know, the most basic common problems and then just kind of build on top of that. And that needs to,
00:39:10
Speaker
happen in parallel with the, with, with, with a rollout strategy. So where do you start? Well, we see successful platform teams usually start from, obviously understand the medium common denominator, but then sort of deploying it into the context of the most advanced teams within the org. Because those will also be the ones that can handle
00:39:36
Speaker
all different levels of obstruction, and you usually want to start by fixing the lower level of obstruction before you obstruct more and more, and are also the ones that maybe are more familiar with all cognitive technologies. They're not scared of looking at a hunt chart or at a Terraform module or something like that.
00:39:56
Speaker
And so that's where we usually see kind of the rollout happen. Then once you figure that out, you start building your first advocates effectively internally, right? Then you start rolling out from there to different teams. And that's, I think, where it becomes really interesting and where
00:40:13
Speaker
this toolbox really kicks in because it's really about, and also the feedback loop with developers, because then it's really about, hey, okay, I might have a senior backend engineer over here that is really comfortable with the YAML files and really enjoys all sort of like jubectl, whatever.
00:40:38
Speaker
And for them, if I were to, you know, let's say introduce a very, you know, UI based sort of like click ops thing, they would hate it. They would hate it. They would totally push back.
00:40:54
Speaker
They would never use it. And that is honestly one of the, it's like, I think is the number one reason that we see platform initiatives fail is ClickOps, basically, is somebody sitting there thinking, hey, I'll just build this pretty thing on top of whatever my delivery setup is. And I'll just roll it out and everybody's going to love it. And management loves it because management always loves it.
00:41:21
Speaker
looking at a dashboard and see, oh, how are my door matches doing? But that's not the case for developers. And while that might make sense for, let's say, the flip side of that, which is your junior front end that doesn't care whether you're running Kubernetes or somewhere else, they just want to deploy a change to some environment and see it working.
00:41:51
Speaker
The problem is if you optimize for this, everyone else is going to hate you. And usually your senior back ends are the ones that have more of a stay in the organization. And so it's really important that you have very flexible golden paths.
00:42:07
Speaker
And that developers, so to say, can choose their own adventure, right? Obviously within the constraints of your rules and your RPAC and everything that you can do with the platform to have an auditability and make sure that, you know, things don't screw up all the time. But it's really important that you listen to developers and that you
00:42:30
Speaker
build a seamless way of integrating into their workbook, which is why we see stuff like, you know, score in the community really pick up. Yeah, because it's a, it's a, it's a, you know, a fully code based way of going about your application configurations that is not, you know, that is environment agnostic, and that is not
00:42:54
Speaker
forcing you into, hey, you know, you really like, uh, like, you know, thinking about your configurations as code and here and there. Oh, now you need to jump into this UI over here to do something. And then you need to jump back in somewhere else. People just hate it.
00:43:11
Speaker
Yeah, I mean, that's a good kind of point in which I have a kind of a specific question, which is more or less like you're so you're treating your platform as a product, you've identified an advanced team within your organization to kind of start with. And so you're building sort of an MVP part of this, this platform that's sort of based on this team, right? So it may be built in a way that satisfies, you know, the 90% of
00:43:38
Speaker
use cases for that one team. As you mentioned, you have a number of other teams that may use different tool chains or have a different path. Now, a question on product adoption. If you've built this internal product,
00:43:53
Speaker
And you have other teams that you now want to say, okay, well, you know, this advanced team is now say, you know, happy, they're onboarded, they're on, you know, on the same level, we're communicating well, how do you get sort of these other teams to then start using the platform as well? Right? Is it, you know, what changes do they have to undertake? Or do they make sort of a new golden path through the platform with you?
00:44:16
Speaker
How does that work? Correct. It's the latter. You just design your goal for different types of users. And sometimes those users will be, you know, different types of users within the same team. Sometimes those will just be different teams entirely. Right. And again,
00:44:35
Speaker
star, I think from the sort of like lower common denominator and lower level of destruction. And then if it's needed, go higher and higher. And that is needed in very large organizations. The smaller, the more advanced you are, the less is something like this is needed. We see
00:44:55
Speaker
internal developer platforms getting a lot of traction at very large enterprise where, again, the problem is, hey, I'm a developer, I just want to do this simple thing and I just can't. There's just no way, it takes me 10 days, whatever, right? And so at that point, rolling out a platform, everyone is just so happy, right? Because it's a step change better for developers, operations, management, and so on. Usually the champion remains kind of like the operations, you know, stressed out,
00:45:24
Speaker
I'm a bottleneck type of thing. But we've had a lot of requests also from engineering managers that were frustrated because like, hey, I can't, my team cannot operate, right? But usually the champions is on the operations slash platform team side of things.
00:45:43
Speaker
Now, when you go to smaller teams, let's say now you're like 100, so you could benefit from an internal development platform, but you're 100, all pretty advanced, all pretty familiar with Helm charts. Then at that point, of course, the pain points are not as badly felt.
00:46:05
Speaker
It's very likely that you still have two, three people in the org that are under extreme pressure. I don't know if I'm going to survive today, but for the rest, it's a bit of a tragedy of the commons where for the org would still be a net positive.
00:46:24
Speaker
for the selfish perspective of the developer is, right now I just like this person and now I need to get used to this other new workflow. It's all right. But that again is why you really want to make sure that you listen to them and you just add extra value at every step of the way. And you really don't interrupt the workflows with some weird UI.
00:46:51
Speaker
Okay. Gotcha. So, uh, I think my next question is more around the platform itself. Like, okay. I was listening to a talk that one of your peers gave at KubeCon or it was on the CNCF YouTube channel about building these platforms, right?
Static vs. Dynamic Developer Platforms
00:47:05
Speaker
And he mentioned something about building a static development platform, internal development platform or IDP and a dynamic one. So like we have covered about how the importance of that platform, but what are the different types, like how, and how they can help, uh, an organization.
00:47:21
Speaker
Yeah, that's a great question. What we see in the industry is most delivery setups today are static. Also, most internal developer platforms that we see today in the market are static, which means they do a great job at enabling developers to deploy
00:47:40
Speaker
an updated image from one stage to another, as long as the infrastructure and the dependencies of the application don't change. So if you have this very simple image update use case, which is, to be clear, with the majority of cases, then everything is good.
00:48:04
Speaker
But the static setups tend to break or at least cause a lot of overhead for all the teams involved. If, you know, let's say you want to start, you want to do rollbacks or you want to do, you want to change conflicts. You want to add new services with new dependencies. You want to refactor. You want to change the infrastructure. You want to spin up in your environment, right? Or you also just want to like audit the delivery flows and do like a hot fix to production and something like that, right? Then the static.
00:48:33
Speaker
world tends to break.
00:48:38
Speaker
It breaks because if you want to just do like a little change, right? Now you need to go and do changes across all these different config files of which nobody really has a good overview, right? And if you multiply that times, hey, I have like N, you know, like X config files per environment, and I have Y environments, and I need to do like Z changes every day, like,
00:49:06
Speaker
You get thousands of new versions of configuration files, depending on your scale, on a weekly basis. It becomes really, really hard. What we advocate for, specifically as HumineTech, I should say, rather than the overall platform engineering community, is what we call dynamic configuration management.
00:49:29
Speaker
which is a methodology to structure the configuration of compute workloads, where developers create workload specifications describing everything that their workloads need to run successfully on an environment agnostic basis.
00:49:46
Speaker
And then the specification is used to dynamically generate config files and that is from the platform orchestrator so that will be human tech. And then those config files are deployed and they're used to deploy the workload in the specific environment.
00:50:03
Speaker
Right. And so in this dynamic configuration management world, developers don't need to sort of like define developers or ops don't need to define or maintain any environment specific config file for their work. Right. So it's kind of like configure once, you know, deploy everywhere and
00:50:23
Speaker
And the way we propose people do it is through a combination of score, which is this open source workload specification that lets you do that environment agnostic sort of like, hey, here's what my workload needs to run. And then once you deploy, the platform orchestrator will worry about, it will automatically spit out the right dependencies, the right permission, respecting the right permissions and so on across the different
00:50:53
Speaker
Yeah, I think this ties back to an episode that we did earlier this year with Cluerel.sh, where they also have a platform where you can bring in new versions of your application code, but then what their platform enables is you can select the deployment infrastructure or deployment environment as well. So let's say I want to try out with a specific version of AWS EKS or a specific version of GKE.
00:51:16
Speaker
I think they dynamically generate all the terraform files and all the cloud specific configuration files that you might need to deploy your application so i think that.
00:51:27
Speaker
We already see kind of a trend moving towards this dynamic platform because as you said, right, static platforms don't scale well. And then that X by a cross Y cross C doesn't really add up. So yeah. Yeah. I think that's a really good point. Yeah. And we do see players that, you know, usually the thing is they usually focus on either an infrastructure orchestration side of things or on the application configuration side of things.
00:51:50
Speaker
We haven't seen something that sort of like does both at the same time and that's kind of like what we push. But yeah, like the trend is towards dynamic and that's really healthy because static is just a nightmare. Agreed, agreed. You might as well go back to waterfall, I guess.
00:52:06
Speaker
I don't know if you go that far. That's a whole different nightmare. Yeah, someone once told me abstractions, abstractions, abstractions. I can't say that apparently three times in a row. I guess next question, maybe a good follow up is, some DevOps engineer out there might be thinking,
00:52:32
Speaker
You know, I just learned this DevOps thing. I just got my CK. I finally feel at home with what I'm doing and now Platform Engineering is out. You know, are there new skill sets to interact with said platform? You know, you said Humanitex, one of them that there's, you know, bound to meet others out there. Are there new skill sets that are useful and or are others applicable? Like if I already know Kubernetes well or GitOps or those kind of things, what is applicable and maybe what's new?
00:53:04
Speaker
Yeah, I mean, honestly, no. Like if you are familiar with IC, GitOps, that's everything that you should be familiar with anyways, right? In the cloud native world. I really think it's not about the technology. It's really about the skills, the communication skills that we talked about and the product mindset.
00:53:26
Speaker
just being willing to listen and have some technical empathy, I think is really more of a mindset shift and of a new set of soft skills, rather than like, hey, you need to go get this certification. Otherwise, you can't be a platform engineer. Now, that said, that is something that we are thinking about as a community, is
00:53:53
Speaker
Because I feel like the platform engineering community has spent the last year and a half trying to hammering into people's heads, hey, first of all, listen to us. And then second, if you want true, you build it, you run it. If you want true DevOps, then the way you do it is through a platform. And for a while, nobody was listening. Now it feels like a few more people are listening than before.
00:54:19
Speaker
And so I feel like now there is a situation where people are like, all right, I'm in, right? This is hot. I heard a lot from engineering a few times, like, where do I go? Where do I start? And then it's crickets, to be honest.
00:54:36
Speaker
There's just not enough good starting points for people out there. We're talking still, I think, high level on this thought leadership sort of layer around, hey, why does this matter? Blah, blah. But it's like, all right, where do I go? How do I get started? That's a great question. Why don't we try to answer that here? How do we make this real? What's out there for people today?
00:54:58
Speaker
Yeah. And so there is not too much I think
Demand for Platform Blueprints
00:55:02
Speaker
out there. This is something that at the moment, but it's something that I personally and the community at large are really focusing on this year. So we see a huge demand in platform blueprints, right? So like real production examples of IDP setups. Simple.
00:55:20
Speaker
In June, we're running the second-ever platform con, and we are going to have a dedicated track just for that. I'm personally looking forward to see some really great talks around that from platform engineering practitioners that will really in detail walk through their setups and hard learnings how to get there. If you could give a elevator pitch for what is a platform blueprint, what would that be?
00:55:48
Speaker
Yeah, so I think the TLDR is something that I can look at, I can immediately understand how this works, how all the pieces are connected to each other, and I can understand how that would potentially transpose to my setup. That's it.
00:56:09
Speaker
Right. It's something that like, Oh yeah, I get it. And I also get how this might work for me. Right. And obviously like maybe you need to swap GCP for AWS or whatever, but like you understand how that would work. And so the talks are going to be one thing. And then we personally, as you might like, are going to open source a couple of reference architectures as well.
00:56:29
Speaker
on interndeveloperplatform.org, where we're also pushing other people to do so. I really believe that should become the main reason why people go to interndeveloperplatform.org. Until now, it's been, again, around thought leadership and about definitions and like, what is an IDP? Why an IDP? What are the core components? What is the tooling out there? And I think now it's about, all right, let's take that tooling and let's start bundling into real reference architectures that people can use and just fork, and then go from there.
00:56:59
Speaker
You brought a platform con, right? You said it's in June, but where is it? Is registration open? How can people register for it?
00:57:07
Speaker
Yeah. So it's completely free. It's virtual as platformcon.com. Everybody can sign up from all over the world. We had 7,000 signups last year. We're thinking 15 to 20,000 this year.
PlatformCon 2023 Announcement
00:57:20
Speaker
And there's just a lot of interests. A lot of people that were talking about it last year really made a lot of noise in the industry. And I think for a very good reason, because it's a conversation that we really need to have. And we have some amazing
00:57:34
Speaker
just practitioners, DevOps leaders, and platform practitioners that have been building these platforms for the last five or ten years. It's great. I'm personally really excited about the format. I feel like a lot of virtual events just take an offline format online, which basically ends up
00:57:56
Speaker
being a eight hours long stream of twenty five minutes back to back talks and everybody wants to die and instead what we will we propose is hey i am basically hosting a. Kickoff every like every day it's two days conference twice a day so one for the morning what is morning.
00:58:16
Speaker
And then off you go. Go consume your content, you know, take your kids to school in the meantime, work, whatever. And then we crunch all the action on this Slack, which is already has like over 9,000 members. And that really, it's just so fun because you basically see
00:58:34
Speaker
The speakers for that day posting a link to their talk and be like, hey, here's the key takeaways, ask me anything. And what that does is it starts like tens of different threads all at the same time. And this is a frenzy of really good conversation that I thought would end last year after the two days. It actually keeps going for weeks because people keep consuming content. The speakers are all just hanging out on Slack anyways. So it's great. I'm really excited about it in case you couldn't tell.
00:59:04
Speaker
I think Ryan and I both have participated in virtual sponsor booths as well, and those are not fun. Those are not fun. We all have, and they're not fun. Yeah, so thanks for helping change the norm, I guess, for virtual events. I think there's a lot of value in
Episode Conclusion and Farewell
00:59:19
Speaker
them, right? Being able to attend sessions like that from home does have a lot of value. So, I mean, that's exciting. And I want to say, you know, we're going to wrap things up here, and I, you know, we'll put links into PlatformCon.com.
00:59:31
Speaker
the platform Slack, any other reference architectures or places people can get involved, Luca, where would you send them?
00:59:38
Speaker
Just PlatformEngineer.org, I think, is the other main one where you'll see all the upcoming events. It's really the home of the community. You'll see all the upcoming events there. You'll see a section on platform tooling, similar to internal or platform.org. You'll see there's a job work as well. We have a store with some fun merch, so people can start from there as well.
01:00:02
Speaker
Right, great. Well, Luca, it's been a pleasure. We're at the end of our time here today, but I think I really enjoyed the conversation. We covered a lot in a short period of time. There's probably a lot more we could go into, so maybe we'll have you back at a later date to see where this community has grown to. But in the meantime, thanks for joining.
01:00:20
Speaker
Yeah. Thank you for having me. I'd love to be back. All right, Bob and platform engineering, a hot topic end of last year. If you're in this community, uh, you know, Kubernetes community in general, you've probably heard the term. Um, and if you were at KubeCon, you've definitely heard the term. I feel like it blew up fast. So this was a conversation I was looking forward to just to dive a little deeper with Luca. Um, you know, what were some of your takeaways from the conversation?
01:00:46
Speaker
No, I think, as you said, right, that platform engineering has been that growing buzzword in our ecosystem, especially. But only a few people understand what it is. And they're still trying to figure out the difference between DevOps and platform engineering and how your team or your role fits into your organization strategy.
01:01:04
Speaker
I think I really like the way Luca basically peels down in. Talking about the need, you can't expect if you're hiring 1,000 engineers a year, you can't expect everybody to know the same set of tools, and you can't wait a couple of quarters for them to be productive when you're hiring so much. Having a platform
01:01:25
Speaker
really solves that issue where you, you have a way of doing things like pushing, writing code, pushing applications to production, things like that. So having that platform really helps. But then the second layer of it, right? Like if you are a team of less than 20 people, if you are a small organization, doesn't really, you shouldn't be thinking about platform engineering because that will add a lot of overhead, then giving you those benefits that we discussed during the episode.
01:01:48
Speaker
So I think the differentiation between like, if you are more than 50 people, 50 engineers, then maybe it makes sense to look at platform engineering, build that team of platform engineers and have maybe an internal product manager who's responsible for building this platform. But yeah, having that understanding of when people should be considering this as part of their organizational growth, it was important to me. So yeah, that was one of my big key takeaways.
01:02:13
Speaker
Yeah, and to your point, I feel like not expecting everyone to know the same. Luca mentioned common denominators when thinking about building a platform. And there's always going to be gray areas. And I feel like those gray areas of overlap in DevOps sort of groups are probably going to relate to some of those common denominators. To how you deliver software, it's probably still going to be a container image, at least in today's world and today's development, cloud native development.
01:02:43
Speaker
There's going to be those commonalities, and that's definitely going to be a key part of it. On top of that, we talked a lot about communication, being leveled up when thinking about platform, which makes a lot of sense, I feel like, because you're going from these
01:03:00
Speaker
separate teams doing development and operations in their own stack and with their own tools where you still need good communication, don't get me wrong, but it doesn't have to boil to a platform team. Now you have a platform team that needs to understand even more across those teams and how to make sure you're not trying to change their life and their world in their every day, but work with them. He said, add value at every step.
01:03:29
Speaker
So communicating how to do that, I could feel like you're easily, if you're not communicating, you're easily going to become the scapegoat who's telling them to do something a certain way. And then just the idea built on that, which is the golden pass, right? You know, I was, you know, part of a team in 2014 that, you know, had a platform even back then.
01:03:57
Speaker
Even though we were a DevOps engineering team, we sort of layered a platform over it. And you did have all these different teams who wind up having very different requirements, very different ways of doing things. But in the end, they're still submitting something to a pipeline. They're still producing a container. But there wasn't as much focus on finding their golden path. It was like, use my golden path, right?
01:04:22
Speaker
So I like that term, that concept. I'd like to sort of double click on that because I could easily become just quite another buzzword or buzz phrase that is overused. But yeah, I'd like to double click on that sometime. And then sort of the concept of let's stop talking about this as a good idea. I think as a community, we've come to realize this is something that's slightly different or an evolution of DevOps that we need.
01:04:49
Speaker
and our want and are already doing. So how do we make it real, right? I like that push at the end that we talked about, like let's get real things going, the blueprints to reference architectures, put things on paper.
01:05:02
Speaker
that people can actually use, and I'm excited to see where that goes in 2023. I know, and again, just following that thought, right? Projects like Backstage that Luca mentioned, those have been popping up, like Spotify built this internal development platform, they donated this as an open source project to the CNCF. Again, they do have a few premium plugins, but again, the entire project is open source. If you want to just take Backstage and build on top of it,
01:05:29
Speaker
That would be a really cool way to get started. So if you're looking at it, check out the project. We'll have a link in the show notes as well. Yeah, and there's a lot of other ones out there. I know backstage is one of them. There's a couple of them out there. If you have here a listener who's maybe working with one of these platforms, we'd love to come on and just have you talk about your experience with it or how you use it day to day. That'd be a great conversation to get sort of the other side of that story to really see someone practicing what we're preaching here.
01:05:58
Speaker
Yeah. And, uh, I think that wraps today's episode up. And, um, with that, uh, like one quick shout out, right? Like, I think we now have a YouTube channel with a few, uh, videos on. Yes, that's right. So if you're listening on YouTube, right? One thing we'll appreciate is hit subscribe, hit that like button.
01:06:17
Speaker
We're trying to play with the YouTube algorithms to get us more listeners. Obviously, for all the people that have been listening to audio podcasts, thank you so much. But now we have a new format, so we need to grow our audience there as well. Even if you're listening to an Apple podcast, go to our YouTube channel, hit that subscribe button or like button and help us grow.
01:06:33
Speaker
Yeah, sure does help us grow. And that's a good point, Bob. And it's a new venture for us to bear with us. Also let us know what, you know, any improvements or what you like, don't like. We always like to hear that. And that's, you know, a lot of people have no problem doing that on YouTube. So here they come. All right. Well, that brings us to the end of today's episode. I'm Ryan. I'm Bobbin. And thanks for listening to another episode of Cover Daddy Spites.
01:07:02
Speaker
Thank you for listening to the Kubernetes Bites Podcast.