 We were really thrilled today to have Bob Johnson from Humana join us alongside Alois Reibauer from Dynatrace to talk about what running at scale looks like and what that means for a company the size of Humana. Come on up. Thank you Bob for joining us. Thank you. Nice shirt change. I like it. For those of you that haven't checked out the new shirt, move fast and don't break things. I'd like that's a good theme for I don't know maybe the last week for sure. Oh yeah. Sorry Facebook joke. Too soon. Too soon. But thank you both for joining us and I was really excited to talk to both of you because I do think for many of the users particularly here the work you're doing at Humana is super interesting and super important because we talked a lot about some of the other industries changing but health care is changing really quickly as well and digital transformation is obviously something you probably think a lot about these days and and it's been great you know we were talking a little bit earlier and you're like well we're not running that much but when I think about enterprises even if you're running just a few apps you're still running at scale because you still have thousands and thousands and thousands of users even if it's an internal application and that creates a whole different set of issues. So you tell us a little bit about the work that you're doing and really what your focus has been. Yeah thanks Abby absolutely. So we're no different from any other customer obviously in terms of wanting to go fast and and really doing it at scale focusing on pipelines and and delivery but we absolutely bring in a component of operational excellence and focus relative to operations to the table and specifically what can we do to make sure that we stay as resilient and as stable as possible and there are lots of components to that. The other piece of this is really when you think about the DevOps three ways the middle way being feedback loops which we've already talked about a couple times today tightening those feedback loops is of utmost importance from from our perspective. So we really look at a lot of different ways to do that we employ several different tools and several different attempts to go after that. Some of the things we're able to do is bring some of our legacy monitoring solutions to the table application performance solutions application performance management and you know Dynatrace is really a key piece of that as well but it's really all about what can we do to optimize our processes and and really tighten the feedback loop. How can we get as much information back in the hands of the developers as you know as we possibly can and from my perspective that's the most important thing that we're doing to really tighten things up. So just to hit on that point because I think it's important the feedback loop is the most important part. Absolutely so it's it's very important to go fast but again don't don't break things right that that's really a key thing and then the other piece of it I think is you know what can we do to employ shortcuts so a lot of IT is really focused on what can we do to drive automation and to optimize things that are sort of traditional IT. So how do we you know really take a quantum leap forward so if you think about the Boston marathon it's a 26.2 mile race think about if you can actually you know cut that kind of kind of in the middle or across the field a bit and save a little bit of time and that's kind of the concept that we look at relative to Cloud Foundry that's a key piece for us again using that for systems where it makes sense we've got several systems where it doesn't you know necessarily apply but that that's really been a key part of our strategy. Which is it's phenomenal and I know that's probably top of mind for everybody else in this room as well. I know Dynatrace has been a big part of that journey for you. Is that an asset that you're going to continue to build on or you know how what kind of difference is that making to your organization? Yeah from our perspective Dynatrace is really a capability that we've invested in for quite some time right so it was very fortuitous for us to be able to to leverage that technology and apply it as a feedback loop specifically to you know kind of to the new world if you will the Cloud Foundry world so having the ability to take an existing tool that we had built a number of operational processes around and apply it in the Cloud Foundry world was really key for us. So we've kind of talked a little about the tech but I feel like where you're also saying is the culture change has been an important piece of that change for humana as well and thinking about how do you make this change how has that journey been for humana? Yeah it's been a journey and it continues to be a journey right so obviously with any any group you've got a number of people that embrace the change that that you're looking to to drive but the desire for speed doesn't change I mean that's getting you know much more important as we as we move forward so I think the the need for acceleration is really key. Again it's like most other things you you roll out certain approaches to software engineering and development and some groups embrace that other groups are a little slower to the table but you know as part of our journey we're really focused on you know leveraging Cloud Foundry to really drive that and push that push that ahead. I'm gonna put you on the spot for a minute I know I told you I wouldn't do that but I am. What is what is one lesson you've learned from this journey so far that you think would be really helpful for other people that are maybe just starting down this path and are just trying to figure out what to do? Yeah that that's actually an easy one. The one thing that that we find has been missing in a lot of enterprise spaces specifically you you have and we all know this you have an operational world and a development world and it's very difficult to kind of find that common language so to the extent that we can leverage capabilities and tools to drive that common language to the extent that we can maybe take some of the tools that that people are using you know in the operational space and shifting those and putting those in the hands of the engineers that's probably the biggest lesson learned is to focus as much on that as possible so again don't just necessarily look at these tools as monitoring tools look at them as performance tools don't necessarily look at them you know for the use case that's maybe the most popular use case but try to find places where it applies to the developer and and I think from my perspective that's that's been our number one lesson learned. That's a powerful lesson and I love how you're you're putting the developer first in the story it's not the the tack first or the end game first it's you're putting the developer first as you build the solution around them and really absolutely make that fast. Yeah while still doing that under control. Well control is helpful yes you know it turns out you if you are a publicly traded company you just can't take the year off to go and rehab everything. Exactly. Because I think that would make it easier though right. It might. But you know it's been such a fascinating to hear that because I think what we hear a lot about is starting that journey but how do you get started and how do you get the momentum and then how do you do it and still able to meet the demands of your customers both internal and external if you're a large company and I think Humana really hits on that because this is a journey you've been on for a couple of years but you know to your point there's still a little road left in that journey. Absolutely start small and then scale which is the name of the game. And don't break things. And don't break things yeah. Anything you would like to add from Dynatrace's standpoint around this journey I know you've been working quite closely with Humana for a while on this. We're working quite closely with Humana and a lot of other enterprises and what we have really seen happening over the last year that cloud really goes mainstream in the enterprise as well. So we have really seen it also back to that statistics that you showed before that it's no longer that pet project inside the company but cloud is a core building block of your future application and strategy. And that's also why we how we see Dynatrace really fitting well together with Cloud Foundry. So the analogy that we use is autonomous self-driving IT and the idea is you want really like a self-driving car. You want it to get you from A to B right? You don't care so much how you get from A to B. You don't care when you have to shift, when you have to change lanes, when you have to change direction but you want it to happen in a safe way and that's how the two technologies play well together. So what we provide is more or less the software behind to tell you okay something's going wrong along the way. You should go somewhere else or it's not so smart to change lanes right now because there's a car coming right after you but what we also tell our customers along the journey we're not running applications. That's not what Dynatrace does. We provide a platform that helps you to manage and optimize your applications but not run them and that's why we tightly work together with Cloud Foundry and also tell our customers in order to really go down that vision of having your IT infrastructure managed and run itself and only get you into the game in the case of any exception. You also need to have an underlying platform that's actually able to do that if you have an application that's your so nothing against legacy technology but a 25 year old GE application that you still deploy via FTP on a Solaris system that's more or less out of warranty right now. It won't help you to have the smartest systems realizing that there's a problem going on because the system is not able to do anything with this information by scaling up by changing things by rolling back changes and so forth. So I just learned something my Solaris system's out of warranty. You can't do it on Solaris. Okay that's gonna be a problem for later but yeah I think that makes sense is really you know how you help that how you package that up but you know one of the things we start talking about is as we're running at scale what does that look like how do you manage that how do you orchestrate that and also you know what happens when there's a failure how do you know there's a failure how do you you know really you know pinpoint that. I think the key difference is you don't manage it yourself you have to really change your approach towards operation so what we've seen in the past that would obviously what we refer to more or less a data on glass approach so you have this big knock room so this room so there's no daylight pretty much like in here. Hey we have daylight this yeah back there. So a lot of screens you watch those screens and then when something happens you kind of act on reactive. You're purely reactive it's a scripted action and you have people watching screens but that doesn't scale it's fine if you run at say 100 servers it's already a lot of work to do with 100 servers don't get me wrong but what we've seen what happens when people approach move towards a more cloud native approach that their environments tend to get 20 times bigger so it's 20 times the size even for the same applications because you break it down into microservices it's more individually-scaled components which is great it just adds complexity. Does it now mean that you will also add 20 times the people? No you don't so you also have to make this part of looking at data analyzing data and automate it as well and that's where really the approach that we have comes in where we have built an AI component and I know today everything is AI but think of it more or less building a system that does a lot of the analysis that a human would do into a software system and automate these processes because as you automate how you run your replication it's also automate how you manage your replication and human intervention should be the exception and not the rule like today if you have to SSH into your pass platform something is terribly wrong right now and it's the same for when you push something through your pipeline if a human has to look at what you're pushing through your pipeline something is most likely wrong as well it should be wrong that should be the reason why human intervention is needed we heard before about shifting left security we see pretty much the same happening for performance it doesn't help you to develop and that at the very end realize once your first customer hits your side oh this thing is actually slow or they cannot log in or they cannot perform any other tasks so that shifting left piece is also part in there and what we've also seen that this additional visibility to Bob's point what we've really seen that visibility also creates responsibility you can't expect your developers to feel responsible for things breaking into production if they have no access to their data how can you feel responsible for something that you don't know about and the positive side is then responsibility creates a lot of entrepreneurial thinking with people okay this is how I can improve it it starts to be really fun seeing okay this change really had that positive impact so I think that's how we have to rethink how we work with data it's not just keeping the lights on as we've heard before it's only how do we move forward and what's the next step for us and don't break things move fast the cautionary tale don't break things you know maybe you should some of those don't mark Zuckerberg well I find myself hilarious but that's also I think a great example here to like that example you should know when things happen and said and those were a way of things coming to play for the enterprise we talked about security also before and especially in the observability domain and the monitoring domain security also plays a vital role so back to that point so like historically we see usually you had like monitoring individual for individual apps so you were looking at apps individually but now you have a platform that runs pretty much full stack a lot of stuff so how do your segment so how do your segment there and how do you ensure that people only see what they should see so I think that security topic is also extending into monitoring as well and we have for example introduced a concept which we call management zones so we're directly extracting organization information and other information out of cloud foundry and dynamically create access rules for for this type of information which is also key because the two of us might be able to look at the same dashboard but we might be only able to see different types of data so yeah that's awesome and I love how you know you're talking about being proactive and and really taking a proactive approach to the way you're architecting and deploy interaction you're tying that with your the way you're really putting that feedback loop in I think combined those two things really put you in a powerful position to not only deploy new apps but scale and for the for the Humana case you run a lot of you have a lot of very critical data like other companies you were choking about in this case even more critically but you're on the one hand you talk about like information democracy making information available which you also want to do for a developers but still you need to ensure that they only see the information that they should be seeing right right well that is a super powerful story and I'm I'm really thankful that you were both able to join us today Bob thank you so much for sharing the story that man has been on I know that everyone is is taking notes about the journey you're on and all of us always thank you thank you for bringing the t-shirt thank you both for joining me