A RedMonk Conversation: A RedMonk Conversation: Zack Chapple on Zephyr Cloud, Micro Frontends, and Module Federation

A RedMonk Conversation: A RedMonk Conversation: Zack Chapple on Zephyr Cloud, Micro Frontends, and Module Federation

Share via Twitter Share via Facebook Share via Linkedin Share via Reddit

In this RedMonk conversation, Zack Chapple, Co-Founder and CEO of Zephyr Cloud, discusses how micro frontends and module federation are empowering developers to work more efficiently. Zack shares insights from Zephyr Cloud’s recent Acceleration Week, the importance of open source, the role of AI in development, and how Zephyr Cloud is positioning itself to meet the challenges of the evolving tech landscape.

This RedMonk conversation is sponsored by Zephyr Cloud.

Links

Transcript

Kate Holterhoff (00:12)
Hello and welcome to this RedMonk conversation. My name is Kate Holterhoff, Senior Analyst at RedMonk, and with me today is Zack Chapple Co-Founder and CEO at Zephyr Cloud. Zack, thanks so much for joining me on the MonkCast.

Zack Chapple (00:23)
Hey, thank you for having me.

Kate Holterhoff (00:25)
All right, so let’s start with an introduction to Zephyr Cloud. What problem do you solve?

Zack Chapple (00:30)
We basically help people get the production at the speed of thought. So I think that’s kind of the easiest way to sum it up.

Kate Holterhoff (00:36)
Okay, well, so for folks who are a little more technical, I mean, MonkCast, we’ve got a mixed technical and business audience here. So I wanna make sure that we’re offering explanations that are gonna make sense to everyone here. So we got the pitch of what it is that Zephyr Cloud does. Talk to me a little bit about brass tacks. What is it that you’re actually doing behind the scenes in the stack?

Zack Chapple (00:55)
Yeah, a lot of what we’re doing is focused around distributed frontends. So teams, as they’re starting to decompose their user experience, want to be able to ship things and visualize things as quick as possible. The moment you get into a federated experience, it starts to get a little bit slow. And when we talk about moving at the speed of thought, we talk about being able to basically compose user experiences in a production or near production environment. And Zephyr allows you to do that by publishing those versions

and of those remotes directly to the customer’s edge. we’re able to, most of our releases at deployments are about 300 to 500 milliseconds. So it’s about as fast as a local host development, except you get to see what you’re working on in production or staging or wherever your company lets you.

Kate Holterhoff (01:39)
All right, and for folks who aren’t familiar with micro frontends, how would you define that?

Zack Chapple (01:44)
So micro frontends are kind of like the way I usually describe it is when you started to decompose monoliths in the backend to microservices. Micro frontend is anything that you want to be able to ship or use at a different velocity than the rest of your platform. Things that are low churn or high churn are things that are prime candidates for micro frontends. That could be something like the header, product listing page or something like that. Headers usually decompose because you don’t deploy it and change it as often. Things like a product listing page, you decompose.

because you change it very often. So anything that you want to decompose.

Kate Holterhoff (02:16)
OK, yeah, and modularity is so important in so many architectural patterns. Does the micro frontend pattern complement other architectural patterns, or is it just kind of doing its own thing?

Zack Chapple (02:29)
It goes really well with like a microservices pattern too. We see a lot of people doing microservices and BFFs, so backends for the front end and then micro frontends. do that across their entire stack. And a lot of architects tend to like that because all the patterns match all the way throughout the entire stack and it makes their diagrams a lot easier.

Kate Holterhoff (02:47)
Yeah, that makes sense. All right, and I speak to lot of front end and mobile developers who are very vocal about the pain points and bottlenecks that they’re encountering in their own workflows here. I mean, just to name a few, the App Store creates a lot of friction. Also, the decision making around whether or not to use PWAs. I’ve written a little bit about this. I’m interested in how Zephyr Cloud is empowering mobile developers.

Zack Chapple (03:10)
Yeah, that’s something we’re going to be talking about during our launch week that hopefully this podcast airs during. So one of the things that we did recently is we worked together with the Callstack team and something called Repack, which allows you to be able to use Module Federation on mobile. And the really exciting thing about that is it allows you to do over the air updates really easily and in a modular fashion. So traditionally with mobile applications, have to use either App Store publishers,

updates to be able to get new version of your application out or use something like CodePush or I think Ionic has one and then Expo EAS like those kind of tools can help you to do over-the-air updates. With Zephyr and Module Federation we can do the same thing except it’s leveraging your infrastructure so if you have a CDN like you know Akamai or CloudFlare or something like that

you can hook that up to Zephyr and now your over-the-air updates are served by your infrastructure rather than ours or some other third party, which for security reasons is kind of a nice thing.

Kate Holterhoff (04:09)
Right, and you’ve mentioned Module Federation a couple times now. For folks who aren’t familiar, what is that?

Zack Chapple (04:15)
So the way I like to talk about Module Federation, I always go to the back end because people know back end analogies a lot more than they know front end analogies it seems like. Module Federation is like Docker for the front end. It’s how we encapsulate different pieces of functionality. A lot of the APIs are actually similar. What you expose, what you consume, as well as some like lifecycle hooks and stuff like that. That’s what Module Federation provides. And it’s just a series of tools on top of your bundler.

Kate Holterhoff (04:40)
And yeah, that ties in really well with what you were saying about microservices being a complementary pattern. OK. So let’s pivot to talk about how front-end developers are encountering Zephyr Cloud. So I’m really interested in how you work with different frameworks. And I’m thinking of single-page applications like React here. So there have been so many innovations in this space. mean, it’s a full-time job to try to keep up with all the things that are.

moving and shifting and different versions, things like that. But many of these frameworks, they’re intended to optimize where compute is happening, the server or the client. And a lot of the moving forward in this space has been focused on how we’re going to reduce the amount of JavaScript that’s ported into these apps. I’m thinking of hot module reloading, which means that instead of requiring a full page reload, you’re providing real time updates to modules in

applications that are actually running. You’ve got the islands architecture is another sort of example of this, which you’re isolating the apps interactive portions, rendering parts in a static HTML. And again, all of this is just intended to reduce the amount of JavaScript that you’re sending back and forth, which is both environmentally harmful, but also causes a lot of performance issues, especially when we think of how folks are encountering your app outside of the United States or folks who have older devices.

So talk to me about how micro frontends and Zephyr Cloud are playing into that conversation.

Zack Chapple (06:05)
Yeah, a couple things. So one, you hit on a really good point around shipping less JavaScript, right? A lot of the different frameworks and technologies and approaches is focused exactly what you’re talking about, like ship less than on the initial load. And then as the user encounters something, you start to lazy load it, right? That’s where kind of like Module Federation and some of the other technologies fit in really well, because we’re able to decompose those things and be able to almost lazy load them as you encounter them.

comes in is that the moment you start to do that at scale, you run into issues around developers are going to develop, pipelines are going to run, and it’s really challenging in that distributed system to understand who did what where and when.

Right? And where Zephyr comes in is all of those micro frontends now become independently versioned, independently deployable. And now you’re able to see as you move through time which of those versions are changing. Right? And that’s kind of where Zephyr fits into that because you can compose these things together in a reliable fashion and be able to roll forward and roll back sub-second. And that makes it a little bit easier to be able to manage in that distributed fashion.

Kate Holterhoff (07:10)
we’ve been talking about these issues in a very abstract way so far, so I’m eager for us to make things a little bit more concrete. What does this look like in practice from the perspective of the businesses that you have helped? mean, Zephyr Cloud, you mentioned that you’re going to be having a launch week here. You have already had an acceleration week. Talk to me about some of the highlights.

Zack Chapple (07:31)
Yeah, so I think.

The acceleration week is a really interesting one. like we just got done with a acceleration week with a pretty big customer. They actually put it on LinkedIn. So Southern Glazers, we went down there for what we called an acceleration week, which is basically like a, it’s a hackathon. But you know, they’re not a big fan of the word hack. So we, we, call it an acceleration week cause nobody wants to be hacked. But basically what we did is we brought a bunch of really awesome open source people together with our team and their team. And, and we basically set their new architecture

across the entire stack. We built proof of concepts and basically built out everything that we’re doing. And what was really interesting is that a lot of enterprises, it takes days, weeks or hours, days, weeks to be able to get something into an environment where you can actually see it. And even longer, if you want to get into like a production like environment so you could see if you actually fix the bug that you’re working on.

And for Southern Glazers, it was a multi-hour, if not multi-week process, depending on which environment you’re trying to go to in approvals. With Zephyr, we were able to turn that into seconds.

And what’s really interesting is that because of how we’re composing environments, you can see if your change actually fixed in production without exposing it to other users in production. And I think that’s the really big impactful part for them is there’s a much lower risk in actually testing something and being able to expose it.

Kate Holterhoff (08:53)
Yeah, and I was fortunate to attend virtually your Acceleration Week and see how this looked in practice here. And I think what really impressed me about what you were showcasing was that these engineers were able to demonstrate bugs and how quickly they were able to mitigate them. And that’s something that really, I think, speaks to the enterprise as something that everyone’s going to encounter. Like, nothing is going to work flawlessly, showing how you’re able to.

approach issues and be able to fix them as quickly as possible. mean, that’s something that is, it made an impression on me for sure.

Zack Chapple (09:28)
Yeah, it was was fun because like we basically made broken builds in production and what was interesting about that is that we said, look at the build broken and then look, let’s go to Dynatrace and find out what the error was. OK, well I know what that is. Let me go roll this little micro frontend back and in the course of rolling it back. It’s not like a Jenkins pipeline roll back the world. It was a single drop down in Zephyr. I refresh the page and the world’s back working again and that was really

interesting for folks to see because it it just felt smooth right you didn’t have to think about are these things gonna work together you know am I rolling back too far am I rolling back too much and I think that really gets to the benefit of micro frontends and Module Federation too because I don’t have to roll back the whole app as a monolith right all the teams that worked on their features and ship their features in their own micro frontends they get to stay out in production

I can just roll back the small broken pieces of my application and be able to take care of the customers, which is really why we’re coding at the end of the day, right? We’re not coding for fun. Some people are, but like we’re not coding for fun. We’re coding for the businesses that are employing us and their customers, right? And if we can increase the speed to customer impact, I think that’s really valuable.

Kate Holterhoff (10:37)
Yeah, I agree. And I think a lot of times when we’re practicing on new technologies or trying to demonstrate the value, we focus on greenfield projects. And so being able to show a brownfield production environment and that you’re able to make these changes quickly and make sure that you’re able to isolate the issue instead of having to go nuclear on it. Yeah, mean, I could just tell that the audience was very impressed.

sort of signaled to me that these were dragons that they had been encountering in the past and that this new architectural pattern, and with Zephyr Cloud’s help, you all were able to approach that in a really novel way, in a way that was making a big impact for them.

Zack Chapple (11:19)
Yeah, and we talked about the whole stack too, which I think was kind of surprising to a lot of folks because usually when we talk about these things, it’s just web, right? But then we did web and mobile and BFF and we were doing these changes and doing hot module reloading, like you mentioned, on lambdas that were running in AWS. And I think that’s scary to a lot of people, but at the same time, it’s kind of cool to see what’s possible. And being able to use these patterns across the entire stack was a really Really fun.

Kate Holterhoff (11:50)
And you had mentioned that these developers were open source developers. What is Zephyr Cloud’s relationship to open source?

Zack Chapple (11:58)
A really good one. So, Dmitriy Shekhovtsov, my CTO and co-founder and I were both part of the Module Federation Core team together with Zack Jackson, who’s the guy who actually created Module Federation. Zack Jackson is one of our founding members and one of our advisors and investors. basically, the vast majority of the core team are part of Zephyr. And so, we care deeply about open source. The company Dmitriy and I come from, Valor was founded in open source, and it was our open source work.

that got us to Zephyr. really, really passionate about the open source community. And we’re going to be actually open sourcing some parts of Zephyr as well, particularly around like cloud providers and bundler integrations. Obviously, the platform itself will never be open source, because we’re a SaaS platform and we’re trying to build a business. But we do want to make sure that we continue to give back to open source, because that’s why we’re here.

Kate Holterhoff (12:44)
I know that’s going to be music to a lot of developers’ ears there. So the reality of enterprise development, it’s very different from what it was five to 10 years ago. And so we can’t get out of this conversation with talking about AI at least a little bit. So I’m curious what you’re seeing around AI and how you’re positioning your clients to meet those challenges.

Zack Chapple (13:04)
Yeah, there’s a couple of things on that. one, you mentioned like we haven’t changed very much in enterprise and it’s funny because most of development industry hasn’t changed. We’re still like basically FTP a zip file somewhere, right? And that’s basically what most of the cloud providers do. They take the zip from your GitHub and then they build it and then they FTP it to their service, right? Where it gets different is with Zephyr and being able to decompose and version things and all that. That gets really interesting. And that fits into the LLM’s conversation too,

Kate Holterhoff (13:14)
Yeah.

Zack Chapple (13:31)
because a lot of people talk to us about generative UI, like how am gonna use generative UI in a reliable way because they’re not comfortable enough to let it just, you know, roam on their user experience because they don’t know what’s gonna happen.

When you start to use generative AI together with Zephyr, now you kind of have that security of being able to version things, understand the impact and be able to roll back much more easily. And then when you can unleash it on smaller pieces of your application at once, it also reduces your risk, right? Because we all know one of the problems with AI is context, right? Being able to keep the entire context of the entire application

in the window is a challenging problem, right? And everyone’s always trying to get larger and larger context. If you can unleash your generative AI on a much smaller portion of your application and be able to iterate on that much quicker, then you’re able to turn that into business impact. And that’s kind of where I see that going, is using tools like Zephyr to decompose the user experience into something that is consumable and iterable on your AI.

and then being able to track the impact of those changes through Zephyr and your other tools to be able to see if it’s really making it better or worse.

Kate Holterhoff (14:45)
I like that. And so Zephyr isn’t planning to incorporate AI into its internal functions in any capacity. Do you guys have a chat bot on your docs? what’s the, you know, are you going that route at all?

Zack Chapple (14:56)

Chatbot, like, eh, we’ll probably have a chatbot. Everyone needs a chatbot. But no, I think the AI aspects of what we’re looking to do is, so there’s a couple things. One is around self-healing infrastructure. So that one’s a really fun one for us because nobody likes PagerDuty and being able to understand what to roll back and when is really important because…

Kate Holterhoff (15:01)
Gotta have a chat bot.

Zack Chapple (15:19)
A lot of the tools for rolling back automatically are not necessarily sophisticated enough to understand if you really should have rolled back. And a real good example of that is I had someone tell me last week that they rolled out a new version of Search for their platform. And it started throwing all kinds of errors. So they rolled it back. What they didn’t know until several days later that it was actually converting at a double digit percentage higher. So they were making more money, but they were also throwing more errors, and they rolled it back.

You might not have cared about those errors. You definitely cared about the double digit growth of your business, right? And for some businesses, that’s really important. For other businesses, they may say we can’t have errors because we’re in compliance or whatever kind of industry that makes sense, right? You always want to roll back. For like e-coms and stuff like that, if you’re converting higher, roll forward, fix the errors. Don’t roll back the thing that’s converting higher, right? And that’s kind of where our self-healing infrastructure will get really interesting. And the other part is around understanding developer productivity and impact.

There’s a lot of things that we can do there. A lot of people are trying to measure that. But with Zephyr, we actually have the ability to do it properly. I would consider it properly. And it’s measuring the things that matter to the business, not things that are arbitrary like pull requests or lines of code changed or story points moved. Those are things that don’t actually matter. People will try to tell you they do, but they don’t. It’s the impact of those changes that really matter, either to the other people on your team or to the customers at the end of the day.

Kate Holterhoff (16:49)
That’s phenomenal. We’re about out of time, but before we go, how could folks hear more from you? What are your preferred social channels and are you or the Zephyr Cloud team planning to participate in any events in 2025?

Zack Chapple (17:00)
so Discord, X YouTube Slack we’re everywhere You know, are we have a discord channel? I’ll make sure we get that into the show notes. I think that would be really helpful Obviously you can find us @ZephyrCloudIO on X We do a lot of stuff on there me personally on there as well And yeah, we’re gonna be I think Render ATL is one of the first conferences We’re gonna be speaking at here in Atlanta, which where we where we both are And I know there’s a couple other conferences on the docket

Kate Holterhoff (17:22)
Yes.

Zack Chapple (17:26)
So I think that one will be the funnest one. It’s one of my favorite conferences. So, you know, we’ll for sure be there and then the rest of the schedule probably get posted on our website.

Kate Holterhoff (17:34)
All right, fair enough. I’ve really enjoyed speaking with you, Zack. Again, my name is Kate Holterhoff, Senior Analyst at RedMonk. If you enjoyed this conversation, please like, subscribe, and review the MonkCast on your podcast platform of choice. If you are watching us on RedMonk’s YouTube channel, please like, subscribe, and engage with us in the comments.

 

No Comments

Leave a Reply

Your email address will not be published. Required fields are marked *