PurePerformance - Dynatrace Perform 2018 with RedHat's Chris Morgan

Episode Date: January 30, 2018

We stepped aside for just a few minutes to learn more about RedHat's OpenShift products with Chris Morgan. We chat about their experience in building integration between Dynatrace and OpenShift, excitement about the conference announcements and a shared distrust of mustard-based barbecue sauce in South Carolina.

It's time for PerfBites. What the f*** is PerfBites? The fourth square meal of the day. Don't bogart the PerfBites. F*** waffles. Microwave ready. Blah-ba-dee-blah. Add nutritional value to your brain.

It's time for PerfBites with your hosts Mark Tomlinson, James Pulley, and Eric Regler. PerfBites. Whatever. Hi, everyone. This is Mark Tomlinson. I'm here at Dynatrace Perform 2018 With my good friend James Pulley And joining us is Chris Morgan

The Technical Director of the OpenShift Relationship and Community and Ecosystem Yeah, more or less More or less, exactly For Red Hat For Red Hat Yes Which is pretty awesome

So thank you, you guys are sponsoring the conference Yeah, we are It's our second year sponsoring it And and we're really happy to be here and see what's changing. Okay, awesome. And before we get started, we do want to declare that both James and Chris are from North Carolina. That's the most important thing we can learn. It's wonderful. It'll be a miracle if everyone can understand this podcast at the end. Yeah. I may have to type up a transcript so that people can understand what y'all are saying. That was more Texan, though.

I think I've got the Texas job. Yeah, that's more Texan. Which is a totally different kind of barbecue. Although senators in North Carolina and South Carolina do actually come with subtitles within the state. Absolutely. And we watch them on television. Absolutely. Absolutely.

Awesome. So, Chris, thanks for joining us. Happy to be here. Absolutely. Absolutely. Awesome. So, Chris, thanks for joining us. Happy to be here. A little bit. Let's talk about OpenShift and Dynatrace and how you guys have been working together integration-wise. So for folks that don't maybe know OpenShift, give us a little primer on how that is. Sure. So OpenShift is Red Hat's container application platform. It should come as no surprise, given the containers run in the Linux kernel, that

Red Hat, being that we have a pretty decent Linux distribution over the years, think of OpenShift as the means of actually managing and orchestrating those containers. It's obviously running in Red Hat Enterprise Linux, but then we use Kubernetes as the basis for that orchestration engine to do everything. And then what we also do with OpenShift is we add a lot of things to kind of make Kubernetes usable for the two primary personas, right? When you look at developers and you look at the operations staff, therefore DevOps, or no ops is one of the themes

that we're hearing at the conference this week. Which I don't believe anything is no ops, because if ops exists, there's ops. Even if there's no one doing it, it's happening, right? Well, it's kind of like serverless, right? It's still running somewhere. It's still running on a server. It's still running somewhere.

But yes, the concept is definitely one that gets pushed. And so we started working with Dynatrace around this technology. For Red Hat, we've been involved with OpenShift for about five years, but we made a switch roughly three years ago, and we were some of the first to contribute into the Kubernetes project when it first came out. Oh, cool.

When Google released it. I didn't know that. Yeah, we've been heavily involved. We're probably the largest contributor to Kubernetes behind Google themselves, and so it's really key to our strategy. Doesn't that follow a trend of Red Hat being large contributors to other projects as well?

Absolutely. I mean, that's the beauty of our model in general, right, is, you know, we look at communities and we kind of insert people into these communities that are best of breed, and then we turn them into enterprise offerings or product, right? So there's a big difference between project and product, right? Yeah. And, you know, the way you could think about it is just like you could go to, if you wanted to, you could go to kernel.org and create your own Linux distribution. However, people come to Redhat

because we have something called Fedora, which is an upstream that takes the kernel and several other projects. And then that's our upstream project, Fedora, that we make a product. We do the same thing in the OpenShift world where we have an upstream project called Origin that takes in Kubernetes and other things, and that becomes the upstream to our commercial OpenShift products. Got it. And where we are engaged with Dynatrace is really interesting.

They had a technology called Ruxit, which was their SaaS-based offering back in the day. Which is now called Dynatrace. Which is now called Dynatrace. Which is now called Dynatrace. That's not confusing at all. Well, it was wise marketing, I would argue. That's right. But you can call it Ruxit.

That's right. And so that was kind of our first engagement point. And they really kind of latched on to the whole concept of microservices and containers and orchestration. And this was about three years ago. About two and a half, yeah. Two and a half, okay. So they were still pretty early to their credit. And then was about three years ago? About two and a half, yeah. Two and a half, okay. So they were still pretty early, you know, to their credit, and then, you know, things have just continued to take off.

So this was part of really the Dynatrace transformation, where they were going through their DevOps transformation internally, adopting technologies from the outside, and went to Red Hat as a best-of-breed provider and said, hey, let's partner up, it sounds like. Yeah, it really was, because, again, they, at the time, and still are, trying to reach that microservices environment and become more than just a tool for operations.

You know, they wanted to be really important to those developers at the time they're developing the applications, right? Yeah. And so there was kind of a nice complementary relationship there when you look at that, right? I mean, Red Hat is not a monitoring tools company. I mean, obviously we do contributions to upstream things like Prometheus and other pieces. But, you know, customers are really religious about their monitoring tool usually.

And so, you know, that's one of the really nice pluggable pieces when you look at an OpenShift architecture that, hey, you know, if a customer wants to use this best of breed for both their APM and their baseline monitoring, then it's great. It just plugs right in. Across all the different environments. Absolutely. So talking a little bit more about what it was like to work with the Dynatrace teams, you're actually building some integrations or at least mutual sharing of information. Before we, as we're getting ready, we're talking about management zones. Sure. Before we, as we're getting ready, we're talking about management zones. Maybe explain a little bit the three parts of that as you're working in OpenShift's taxonomy,

obviously Kubernetes namespaces, and then Dynatrace. Andy was here talking about tags. So did that make sense to you? Because, you know, it sort of did for me. It actually did. So, you know, one of the great things to me about Kubernetes is pretty much everything is a label, which shouldn't come as a surprise since arguably the largest search engine on the planet kind of does that

same thing. And if you use Gmail, it's all label-based. So it should come as no surprise that this orchestration project that they have around containers utilizes labels. So again, with OpenShift, that's one piece. Like I gave the fedora to kernel.org as one piece.

Almost think of Kubernetes as the kernel of OpenShift in many respects. It's pluggable and it's those things. And so that's kind of the real piece of it there. So when you talk about the concept of management zones and where it's going, it's able to understand these labels

that are already existent in the OpenShift environment that it's obviously bringing in from Kubernetes and then just apply them automatically as part of their security policies within a Dynatrace environment. Yeah, and Andy was saying Dynatrace sees everything as a tag, which has been a parallel to either the Kubernetes namespace or the label. Which is another kind of great synergy. And then he brings it all into kind of a meta layer that they can just pick whatever they need to out of that for tagging purposes.

That's right. And so that's another kind of synergy, if you will. I mean, if someone started, say, with an OpenShift or Kubernetes environment, they understand this concept of labels. And so then if they layer in a Dynatrace environment via one agent, then they're able to also now see this and work in something that they understand. Yeah. Yeah. And how was it in your experience getting things to work within the new Dynatrace, aka formerly known as Ruxit? You just mentioned one agent. It's their whole new technology. Is that compared to the old agent, which was like kind of crazy. The one agent was easier for you guys to partner and integrate with?

Yeah, it was, and it's gotten progressively better over time. So one thing that we do in OpenShift, so you've got to look how even containers themselves were adopted. So Red Hat, again, we have a Linux distribution, and one of the things we don't like is when people have to use root.

And if you look at some of the original containers, they're all running with root user, and so they're super privileged. That's not necessarily something that's the most secure, especially if you look at a lot of the different things that have happened. Let's just say Equifax. Yes. Equifax. You don't even have to describe GRC, InfoSec, security data classification. Just say because Equifax. But one of the nice things in this relationship is they kind of understood that.

They picked up early that, look, containers are Linux. There's Linux in the container. There's Linux in the host. And so it's like, hey, if the people that kind of are defining commercial Linux say that this is not a secure way, they did a great job within one agent of actually making it to where it can work at certain levels without having to interrogate the host. So you can get a lot of value. And compromising your security posture or anything.

That's right. And so you can get a lot of value out of one agent without running as a super privileged container. Obviously, to do a full stack, you know, but that becomes a case-by-case basis and left up to the admin of that cluster. There's a cost associated with that in terms of computing resources.

There's also a cost in terms of security exposure to make that decision. So the ability to run it at less than a privilege level and get 95% of what you need and still be secure, I think is a good thing. Yeah. And it's, you know, what are you looking to manage and understand, right? That's the whole thing. I mean, what's beautiful to me is, you know, monitoring for the last 20 years plus has always been about capture, diagnose, and fix, right? And when you look at this relationship, Dynatrace does a great job of kind of capture and diagnose, right?

Well, now they can trigger OpenShift to fix if that's an auto-scale action that needs to happen, if it's a pod evacuation that needs to happen, or if a new service entirely needs to be spun up, right? That's kind of where I see things continuing to progress with what we have. To extend that, I mean, the OpenShift world, do you guys have capability? And we can edit this if you need to. Like even triggering different green-blue deployment states. So if you're rolling back, rolling forward.

Yeah, we support blue-green deployments. All of these things, even through your CICD pipeline, so you know you can deploy an entire Jenkins service within OpenShift itself. You can also integrate. A lot of people also are religious about their CICD and have centralized Jenkins farms in a lot of cases. You can pull in those external things and drive the different stages. But yeah, blue-green deployment and AB deployments are very common things that we have within OpenShift.

And so yes, those kind of triggers definitely exist. Very cool. How easy is it for people to get started? Let's say they were a Dynatrace customer and wanted to start playing around with OpenShift or obviously we know there's some trials in the Dynatrace world but what about OpenShift? Is there ways to get on board? So we offer, I mean if you want to do full container platform that you run and manage yourself, we obviously have different 30 day

kind of subscriptions. Remember we don't have licenses. It's subscriptions. And you get an entitlement. So you can do that, you know, 30, 60 days if you want to. But, you know, if you're a developer and you just kind of want to get going with it, right, we've got a full OpenShift online environment. So with OpenShift Origin, we actually take that upstream project,

and something that's unique about OpenShift for Red Hat is there's three business models. Okay. One is something we have called OpenShift Online, which is a fully multi-tenant down to the container level pay-as-you-go kind of service where there's a starter tier, as it were, that is low to no cost, where you can just kind of play around if you're a developer, right? So that's not necessarily the operations end. Right.

Then we have another business model called OpenShift Dedicated, which is full managed hosting, but the same kind of thing. It's run and managed by Red Hat operations. Still multi-tenant, yeah. But it's single tenant. Single tenant, okay. So imagine you've got Coke and Pepsi. Yeah, yeah.

OpenShift Online, Coke and Pepsi, their users may have containers running on the same host in the same cluster. If you look at OpenShift Dedicated. We definitely need to edit that out. Coke gets their own. Again you look at every ship dedicated. We definitely need to edit that out. Coke, yeah. Again, barbecue, Coke versus Pepsi. Coke gets their own managed toasting environment

and Pepsi would get their own managed toasting environment. Right. And then there's the traditional, you know, subscription that Red Hat

has sold, again, for two decades where you are root. I mean, at the end of the day everything boils down to who's root. Yeah, yeah.

And it's, you know, a customer can deploy that on any footprint. That can be any of our certified public clouds, so all the big three, plus several others. Then any of the hypervisors we support, including VMware ESX, our own Rev, our own OpenStack, and then obviously the traditional bare metal OEMs as well. So what's nice, again, because it's based on Red Hat Enterprise Linux, that's everywhere. And so OpenShift is, in many respects, just a better way to even get more value out of your existing investment with Red Hat Enterprise Linux, which is also an interesting thing when you look at how we work with Dynatrace, because they do a lot of legacy business as well.

And so there's less of a rip and replace and more of a true transformation. Absolutely. That can occur. Absolutely. And I'm wondering if there is, you know, in the, what was the first name of the business model? OpenShift Online. OpenShift Online. If there couldn't be, like, pre-installed Dynatrace One agent. So we actually, going back to the super privileged container,

so the Dynatrace One agent can work with OpenShift Online. Install it. Right. As an individual developer, there's something there. I mean, that's going to continue to progress and get better through the concept of service catalog and service broker that we have available now to make things even easier for customers and partners alike. Awesome. That is great. And people can sign up.

They can just go to the website. OpenShift.com. OpenShift.com. Just jump on board. Chris, nice to meet you, man. Nice to meet you, too. Thank you very much. Thanks for having me here. And we'll seeift.com. Just jump on board. Chris, nice to meet you, man. Nice to meet you, too. Thank you very much. Thanks, guys, for having me here.

And we'll see you on main stage? Yeah, tomorrow. Tomorrow? Tomorrow afternoon. And then we've got something here at the partner day. At the partner day. Awesome. Yeah, it should be fun.

Thank you, man. Yeah, man. No problem. Thank you. Thank you.