 Hi everybody, my name is Turk. I work for Dynatrace and I'm here today to talk to you about what are challenges Do we have right now with the open-sec environments not what we have it up and running and I want to show you like what are the challenges and what you can do about it to get on top of your let's say application environment back again so Okay Again, my name is Turk. I work for Dynatrace You see my email address and my Twitter handle if you want to reach out for any questions that you're having Or if you're interested in finding out what we are actually doing with that space listen closely to my presentation And if you have any further up questions, our booth is right behind you if you turn around right now You can see it and yeah, so what Dynatrace actually does we are an application performance Monitoring company we're a world market leader at that and let's say with the rise of the dynamic container and cloud Environment throughout the last years we just saw the necessity to extend our support for those platforms as well Just to be able to give you a full-stack monitoring approach to your applications again So I want to start with a little anecdote how I started with open-stack and how what like my first touch point with open-stack was I Started setting up as three note cluster by hand Where actually everybody tells you not to do that anymore So if you want to run it in production or something like that you rely on let's say open-stack distributions And I tried out some different conveys configurations And what I found out during this process is that troubleshooting and let's say getting a grasp of what's going on in your cloud platform It's actually pretty hard. So it's not that easy. You have to browse through several log files So if you have let's say a medium-sized installation with H a you probably have more than 100 log files That you need to look into if you want to figure out what actually going wrong so I was wondering how actually Companies that use that in large scale worked it in production and I set out to find let's say what are the monitoring solutions out there and Yeah, let's see what I found and let's see what we came up with so Let's take a second to talk about the challenges. We were facing today So we are past day one where setting up was the problem and getting open-stack running and deciding if it's okay to use it or not and and Now the cases that actually companies and and and dev of guys and all of the people that are involved with open-stack and the application our interest in insights into let's say what's how's my application doing and also into insights how's the cloud platform actually doing and You see that open-stack itself actually is an application based on microservice architecture so it's at several microservice on different nodes that are talking to each other and So there's a complexity in the let's say cloud from platform itself So while on the one hand you get Better control of your cloud platform. Let's say in comparison to AWS You actually get insights into what's going on in your cloud platform You on the other hand also have to fix it if something goes wrong. So that's like the two sides of the coin you get with running your own private private cloud platform and The challenges that we are seeing actually is like dynamic environments So dynamic application environments on top of open-stack and like dynamic auto-scaling with heat for example on an open-stack cluster level and Well large-scale right because if you have like a common monitoring solution You need to install manually and you have ten thousands of nodes. You're probably grow old before you even Are able to get the first data So what about the monitoring part of it then? As we've heard we have dynamic deployments dynamic applications that like in the morning are just you have a few instances And then at noon like there's a load peak and you just need to scale your application up Just to be able to handle the load and then in the evening. You will probably just scale it down again So just to save resources and money in that case and you would need a solution that actually supports these kind of use cases, right? and The next thing actually large scale So what are you doing with with environments where you have three thousand nodes or ten thousand services? You need a solution that actually like say takes the complexity out of the system again and provides you with data and answers so There's actually two approaches you can follow so if you're just interested in how your cloud platform is doing you could just monitor the open-stack platform and Monitor compute nodes controller services and see if they're doing fine. That's okay But usually you run something on top of open-stack, right? Usually you run applications on top that are generating revenue in somehow or providing a service to customers or end users So what you actually also need to focus and worry about is all of the complexity that that is actually working on top So you have different application technologies Dynamic container solutions like how cloud foundry open shift given it is and you would also need to worry about the end user performance of your application Because the end user is actually the one that generates the revenue and you need to take care of that and Actually, you need to know if something's going wrong with your application so Again, if you only focus on one thing if you only focus on the open-stack thing this would actually Look reasonable to do right so he's changing the tire because it's a flap Sounds great. So What happens if you don't see the big picture? Well, if you don't see the big picture, you're actually missing a lot Because if you only focus on open-stack itself, I think this is not enough information anymore You also need to take care of the applications that are running on top and the end user experience Just to be sure to stay in business nowadays so The actual question is do you have the right focus with your monitoring solution or Yeah, do you have the right focus on the right thing then and Besides that there's also then again the Often-sighted let's say culture will change that's necessary So we've been there with the DevOps movement that there has been a kind of a discrepancy between operators and developers and The blaming game and just pointing finger at each other than who's to blame now because something's going wrong and We have the same thing again with open-stack because you have let's say a next level of operators that are operating the cloud platform which you would not have had before if you were using a private cloud platform like open-stack and You just kind of have to make sure that they are all on the same boat and in the like have the same view again So that if the boat is sinking all of these four guys are going down, right? It's not just the application developers audio operators and yeah, that's the let's say cultural change And you need to approach that and you need to take care that all of your people that working with that are actually aware of that Now we talked about the focus and We talked about the people and now actually given this facts you should ask yourself. Do you have the right tool for the job? Do you have the right tool for the job that you're actually trying to do you want to? Monitor your open-stack platform you want to monitor your applications you want to get insights into end user experience so What you actually need is Something that kind of consolidates all of these information into one product That you don't have any selective perception anymore. Of course, there are more sites to the same thing as we see here So the operators may see this way and they have a lot at the application developers may see it another way But there is only one truth. So what would be the perfect approach to that actually? Just using one tool that provides you with all the information and just like say Displaced it in another way Just let's say focused on the part that you are interested in and do that you are mainly working on so What we actually need is a tool that provides you an holistic overview of your cloud platform of your applications of your user experience and Fact is Ellen Clark from Susie and Mark Kulay in open in Austin already told us that in a few years the cloud platform will only just be there It's gonna be the defector standard So you actually can then focus back again on your applications and making your applications run smoothly and in order to facilitate that Let's say that monitoring need that there is apparently You have several solutions that are out there right now So I just managed to write down let's say a few categories So you have the log management area that only analyzes what's going on in the log files because open stack obviously writes a lot of log files Like the elk stack Then you have system monitoring tools that mainly focus on resource utilization and if my CPU Utilization is too high and if I'm running out of memory and stuff like that You have some combined approaches even like monaska which kind of like combines these two approaches into one But then again, what about the applications and the user experience? There right now is Well, there is one tool out there that actually does that and I'd like to introduce you to that it's actually We are now extending our support of application monitoring also to the cloud platforms We have been supporting public cloud like platforms like AWS Azure and stuff like that a long time now we're actually like supporting also private cloud platforms like open stack and I would just want to give you a brief intro about what we're doing there So what you would get you would actually get an overview of your open stack cluster like how many computers do I have? How many virtual machines are running on them? how my controller services are doing and what we do we actually like watch all of those services and all of these processes and Create the baseline out of these processes So we don't follow the approach of setting alerts and thresholds We just learn how the processes and services normally operate and then if it deviates too much from the baseline that we've learned We just provide you an alert then so we just want to reduce the spam flood of Let's say other tools out there that provide an alert each time the CPU utilization goes above 60% for example It could be that this is a normal behavior because on Monday morning all of your users use the application at the same time We are able to learn that behavior and know that this is just normal Then at the data at the more detailed level on the hypervisor level You would get an overview of all of the virtual machines that are running on your hypervisor and in addition to that you would get The resource utilization of each virtual machine on the hypervisor So you just get a clear picture, okay I have a hypervisor with eight VCPUs how many VCPUs are the individual VMs using right now Or do I need to move a workload to a different compute node? We actually provide data about CPU usage memory usage network usage and also let's say storage usage use disk space Then again, we support the multi-tenancy of OpenStack as well So you just have if you have several projects for your business units, for example, this is a common scenario You would just like getting an insight into all of the projects that you're getting and you would just be able to get Insights into that from a project view not from a hypervisor view which Could be suitable if you need that and Then again since we're around since more than 10 years We're actually good at correlating stuff and finding out the root cause of stuff so we not only provide you with data we actually provide you with a root cause and we provide you with What is actually the the unique root cause of what has been going wrong in my environment and We provide you with let's say a timeline how the problem evolved over time and all of the components that were acting with each other to let's say be part of the problem and well, you could actually Reboot your problem in in retrospect just to see how it evolved over time and how it got solved so If you want to know more about Dynatrace and what we're doing in the OpenStack environment and what we're doing with application performing at the whole please stop by at our booth and That would be it from my side. Thank you very much