Making the Right Thing Easy with Jon Ravenscraft and Nick Kuhn of Kroger

Episode 130 · 15:33 · SRT · archive.org · all Tanzu Talk

Hi there, listeners. Thanks for tuning in to another episode of Pivotal Conversations. This week, we talked to John Ravenscraft and Nick Kuhn, cloud engineers at Kroger, and hearing just what their key takeaways are from the CF summit that happened in Philadelphia a couple weeks ago. Now, you're going to hear a dizzying array of technologies and projects and efforts that are all underneath the Cloud Foundry umbrella. And so it may seem a bit overwhelming at points where we don't get to dive into each one of these in great detail. But that's sort of the nature of Cloud Foundry. We also had to redact a portion of the recording because some of the insights that John and Nick found from the event involved another user. And so we didn't want to share out their story without their permission. These are some of the things that are great findings at events like CF summit or SpringOne platform, where the hallway track is certainly a source of valuable information. All right, we're here at Cloud Foundry Summit 2019 in Philadelphia, and I've got with me John Ravenscraft and Nick Kuhn from Kroger. So what are some of your key takeaways and learnings from the summit this year? Hi, I'm Nick. So this is part of the biggest takeaway, I think, is kind of the ongoing kind of merger of Kubernetes and Cloud Foundry. And that's kind of, I think we're starting to see a trend where Kubernetes isn't so hyped as much, and it's kind of becoming mainstream, and lots of good progress with Project Irene. And, you know, just, everyone's kind of moving on with our lives, and we can continue to do what we do without constantly focusing on that. Some great stuff I saw around the ISM project, where we're brokering runtimes between Cloud Foundry and Kubernetes at the same time across multiple clouds. So that was a

pretty good session I saw today, which I think could be very useful in the future. And then also good, another great session on K-Native this morning with the two Bryans, Brian McLean and I think Brian... Friedman. Friedman, yeah. And a funny story that Brian McLean's cousin works with us pretty closely at Kroger on the digital team. So some small world type connections there. But yeah, I think that's kind of the big takeaway is there's another session I didn't get to make it to, but the cube your enthusiasm or kind of what you're doing. Yeah. Good title there. Yeah. Paul Sersowski, he's a punny guy. Yeah. So you can always look forward to the replays. Yeah. I always respect those. Yeah. Yeah. So how about you, John? Yeah. I think one of the most exciting things that I saw out of this conference was the forward progress on the Cloud Native Build Packs. And I think the ability to help our developing team adopt a pattern that's more friendly than, like, managing and curating your own Docker image libraries is an awesome step. And it brings a lot of that build pack goodness that comes from Cloud Foundry that may not have been fully adopted in the past to a much more accessible layer. And it also encourages and creates that environment, even on your local workstation, before you deploy it to whatever cloud provider you're going to. And that's really exciting from my perspective. And then they called out the ISM and also like the multiple, multiple foundation management through apps manager is a really big deal for us. And also the replication of, especially like the MySQL services across foundations is a big deal. Yeah. And I think too, as we continue to talk to some of our peers, we're continuing to see record growth and we see other companies similar size also see growth and we're kind of sharing the same pains and being able to talk to them and be like, hey, you know, did you run into this or

that type of thing? Like, oh yeah, I saw that. And similar experience that we could share. Yeah. I mean, you know, one of the nice things I think about this conference and the way it's evolved is there's a lot of end users talking and a lot of them are running at pretty significant scale. Others are sort of more starting out. So, you know, they're, they're operating kind of Cloud Foundry and Anger and kind of, you know, they're dealing with it in the real world where there's compliance and security and, you know, a bunch of layers that a tech startup maybe doesn't have to deal with. So, you know, them adopting a technology doesn't necessarily provide insight into how to navigate those things. Which kind of brings me to a kind of a question on like the process side. There's a lot of interesting kind of learnings about how folks have, you know, adapted the way that they're working. Anything that kind of jumped out at you from some of the different end users that were speaking? I think, I find it funny because we all kind of saw a lot of similarities of how kind of it evolved there where, you know, you kind of start out with a few foundations and then maybe kind of get overcrowded and you start kind of chunking that up and how you segment workload and kind of achieve like PCI or HIPAA type compliance. I think we all kind of went down similar paths. And then, too, you get to talk to some of the other customers like, you know, we're all kind of struggling with the same problems and figuring it out. I think there's been a lot of good work in automation to help kind of make the process go faster and like reduce toil on everyone. And we definitely seen that in, you know, upgrades as well. You know, at least once a month where, you know, repaving the world and doing that before, you know, without Bosch or Cloud Foundry was just not even, you know, tangible or feasible. So I think just the improving the process with technology and making it the least path of resistance to, you know, do the right things is a good example.

From like a community and experiential perspective, too. The idea that you can make the right thing and the safe thing, the best and easy thing is an awesome improvement from a process perspective. Because when you're giving people that springboard into a production environment, that's stable and it's secure and it's relatively self-healing. And it provides all those tools to just dynamically provision, scale up, down, whatever you need during your high demand usage or low demand usage. That, that empowerment for users is just fantastic. What are you, you know, what can you share about what that's done with inside Kroger in terms of that type of empowerment and the velocity that you're seeing? Yeah. So I mean, I think from our growth perspective, we're probably like close to five years in this journey now. Um, four or five, my son's four and a half. So yeah, that's close enough. Um, and you know, we, we kind of started off, you know, small and then we, you know, we moved, uh, like the first real application was our store locator app. It's like where you go to like find a Kroger store or a Ralph's or our different banners and that worked well. And then I was like, okay, we kind of, we kind of like kind of jumped through all the security, security hurdles to get that done. Like we've built a framework of, you know, getting all the, you know, all the teams lined up to, to make that happen. And as soon as that kind of, we kind of like opened for business, it was just, you know, slow, but steady uptick. And then you kind of get the hockey stick approach, but I think we're more just in like, uh, we're still in, you know, rocket ship mode where the line just keeps going up and up and up and, uh, which is a great problem to have, but you know, that brings new challenges of, all right, well, maybe we need to kind of swing this, this, this group to their own set of foundations because, you know, they're getting pretty big and, you know, we've got to kind of start chunking up things a little bit more that we just never problems.

So we never envisioned we'd have, you know, four or five years ago, but great problems to have. Well, to circle back to some of the things that, that jumped out at you, for example, like, um, you know, the K native, uh, so that's interesting to you. Can you tell me a little bit about why? Well, I think, um, in the demo today, uh, it was like a basic go app or, or some app and, you know, there's all this boilerplate code and then, but, really there's just your individual business logic was just one line and, you know, they were able to reduce that down to, all right, just point to this, you know, you know, project riff where, you know, you kind of pass it some variables, but then you just point it at your one line of business logic and it runs the same, same function as the, you know, an app that was previously like a dozen lines of code. So I think that concept of like reducing that, you know, code that's just the same over and over again and just, just going, you know, like I need, I just really just need to run this one line or this two lines to like a function or an event. That's, I think that's really going to be huge. I think today, even we have a, a lot of customers kind of do that today with, with like full, full-blown applications in Cloud Foundry and just being able to say, Hey, like you don't need to, you know, if you're just running kind of more like a function type service, you can, you know, chunk these apps down into, you know, just functions that spin up when they need to, it's been down and we can save a lot of resources that way as well. And that's especially more important as you go to like a per clock tick consumption model from the cloud providers, like every penny counts. Yeah. And then, you know, let's see the, the MySQL replication, for example, what, you know, what was interesting about that? So as a, as a company, we made some very strategic and kind of limiting decisions when we first implemented Cloud Foundry in our environment. And one of those things was for highly available services to be truly highly available, that means you have the scale past a single foundation.

So all the data tiles that are currently available up to basically very recently did not have like a cross foundation approach where I can have the same apps bound to the same kind of services or a service that's in lockstep with a different foundation in a safe way that when we did like platform upgrades during the middle of the business day, we would not impact customers or lose data or have a managed window. So when you have a business model like ours, where your customers don't want to engage in any day, any time of day or night, and we want to give them the opportunity to do whatever they need, whenever they need it, and however method they want to do it in, all those things don't lead to like business windows of like patching cycles. So if you, if you can't tolerate it in place, upgrade and rolling update and paid and everything else, that just became very painful. So we see like that door opening to now or like a MySQL tile installation that can provide MySQL servers from our customer perspective. It takes that out of the hands of our data services team from having to manage and operate and just acquire resources to do it when we already have the platform infrastructure to run it on. So it's, it opens a lot of doors of opportunity there. Yeah, I mean, and that sort of speaks to kind of the, the approach that you're taking with your internal customers, right? Your developers who like, you know, thinking about their experience and just how to continue to make their experience that much better. Yeah, the growth, the growth we've seen just from like the PAS layer has been literally like they called it the hockey stick, that exponential growth. Again, as soon as you make the right thing, the easy thing, like people love it and it just like takes off like a rocket ship. We expect the same sort of thing. And we've already had conversations around, you know, caching and other data, data layer services that aren't readily consumable at the same scale that the pivotal applications are. And being able to mirror those two together and just, it's going to be a real good launchpad for us going forward. Yeah, I think, I think, and then take that in combination with the work that's being done with the ISM project.

Yeah. We can really kind of, kind of complete the circle, so to speak, on most of our services that aren't on the platform now. And really help deliver value on the, kind of, where there's a lot of pain today right now with some of the other more traditional assets. Toil, I don't know. Yeah, toil. Yeah. Pain, toil. I always got to work down that toil. Yeah. Okay. So, what's on the short list when, you know, you're going to go home today or tomorrow? You know, what's, what's going to be different about next week? I kind of referred to the build pack thing earlier. I'm going to, like, go back and get the pack alpha code downloaded right away and start getting that exposed in our developer community. Because, like I said, that's going to be a significant change. And one of the things I love about our relationship with Pivotal and this whole ecosystem here, the Summit, is the ability to provide feedback that actually makes a difference. And it improves lives not only for us as customers, but other customers as well. Because Nick put it out before, regardless of your, your vertical business, at scale, almost everybody sees the same kind of problems, the same kind of challenges. So, when we can unify that voice and identify the things that we do share as priorities, it helps make the whole community better. So, I'm excited about that, especially. I think, too, digging more into PKS and K-native or PFS will probably be very interesting in the short term. It's just being able to expose some of these new features will be, could be huge. So, and then also, right now, I'm, like, I'm currently deploying PCF on Azure. So, I want to finish that up, but I hope I'll have that done by tomorrow. So, yeah. All right. Well, it sounds like you've got an action-packed week ahead.

Any kind of final thoughts, parting thoughts? Was it the Redding Market in Philly? Oh, yeah. Amazing. It's a national treasure. Yes. Yeah. If you're ever in Philly, you need to check it out, because that's, you go in there, and it's just full of delicious shops of awesome food, and it's kind of mind-blowing. It's a gem. Yes. Yeah. So, I mean, and it's, you know, for folks in our audience who weren't at the conference, the conference center is right across the street from it. So, lots of folks have been kind of sneaking over there for, you know, especially those donuts. Yes. Beware. Yeah. Too many donuts. Yeah. It's, you know, I could see, you know, if you were working downtown, like, you could have a different lunch every day for, you know, many, many weeks before you got bored of that place. Definitely. So, well, thank you so much for joining the Pivotal Conversations podcast, and have a great time with the rest of the conference and beyond. Thank you. Thanks for having us. Thank you.

Tanzu Talk · 🍕 Get more pizza.