Welcome to our podcast series on IT modernization in the public sector under written this week by Dynatrace. I'm your host Wyatt Cash, and in today's episode, we're exploring new and better ways for state and local government agencies and higher education institutions as well to migrate workloads to the cloud using monitoring and observability tools and AI ops platforms. Our guest today is Willie Hicks, public sector chief technology officer at Dynatrace. Willie led IT operations in the financial sector for nearly two decades before joining Dynatrace, where he now helps agencies and institutions leverage automated software intelligence platforms to optimize application performance and security in today's dynamic cloud environment.
Willie, thank you so much for joining us and welcome to the program. Wyatt, thanks for having me. Let me start by asking, what are some of the challenges state and local governments and education institutions are facing right now in trying to understand the dependencies associated with the applications on their networks as they look to migrate to the cloud? Well, I state local, higher ed, they're all facing very similar issues and what we also see across the public sector, agencies are understaffed, often overworked, they have to deal with constant budget uncertainties and constraints, and this is really driving agencies to look more towards the cloud actually as an enabler for digital government and to address even cyber security concerns.
What we see though is most agencies and institutions are just in these initial planning stages of thinking about migrating to the cloud, but one constant we do see is that there's a lack of understanding of the target applications they're looking at moving and the nature of those applications, dependencies, etc. So this is kind of a common thread, a common problem, and it stems from some of those initial challenges, they're just often understaffed and don't have the time and resources to do the analysis. Willie, can you maybe describe the difference between monitoring and observability and what do more modern solutions provide that many existing tools may not today? That's a great question, and this is something I often hear this idea of monitoring and observability being thrown around interchangeably.
Imagine your old school job application, no microservices, no complex, API calls, just a Java app, and even let's say maybe there's a service oriented architecture underneath. Now, even though this might be quote unquote old school, you have some advantages here, like you can use Java's built-in logging functions to output logs, you know, all of it can go to the same directory, you're not worried about multiple places to look, you probably don't have large numbers of instances to contend with. You're probably dealing with only a small number of technologies if more than one, and in this type of setup, monitoring or APM platform is often sufficient. Now, with this type of architecture, it's static.
Once you set up that monitoring, you can kind of forget it, walk away, it collects its metrics. You know, when there's a problem, the team probably has enough visibility to kind of dig into it. This type of monitoring really is not adequate though when you move into the cloud, especially if you're dealing with more cloud native apps, where you can have multiple runtimes, multiple platforms, you start introducing containers, virtual machines, and then the cloud itself, the cloud metrics and so forth that are being generated. Now, you have multiple places to collect data from, multiple technologies you might be working with, so using the older APM-like tools, you have to deal with sometimes multiple agents.
You have to deal with where am I going to pull the logs from, how am I going to collect the cloud metrics and container metrics and so forth, so sometimes you're dealing with two or three different tools and trying to stitch it all together. Also, think about what happens when the customer experiences a problem. You now have to trace this transaction across multiple systems, and this can get really messy. So, this is why a true observability platform needs to come into play because it can simplify this complexity, it can automate the collection of data, automate the monitoring of all of these different sources, and then leverage AI to automate the analysis and the alerting.
Well, that certainly makes a lot of sense, but recognizing the budget constraints most CIOs are still facing. What's the business argument for investing in a new monitoring and observability platform? Well, for me, I like to keep it simple, and we're always being asked to do more with less resources, we're developing and supporting all these new digital government initiatives, we're fighting fires, we're ensuring systems are up and running, our citizens are happy, and we're constantly being asked to do more, but we are not getting the resources, more resources to help with that. This is where observability, and especially the AI powered observability platforms really show their strength, because you can think of what could be accomplished if you had a platform that was constantly learning your application, its behavior, the user behavior, adjusting itself to account for new conditions, and then automatically alerting and finding root cause.
These systems can automatically parse hundreds of thousands, millions, even billions of data points to find root cause where it would take your resources, maybe war rooms days to figure it out. So, that's why agencies could be looking into these platforms like Donatrace so their people can focus on the real work of supporting agencies and the customer. Talk a little bit more about the concept of AI ops solutions, how are they different from many existing monitoring systems, and how might state and local agencies and education institutions benefit from using those? Well, to really understand the difference between AI ops and just monitoring, let's think of it this way.
When I talked about observability earlier, we discussed how observability is about monitoring these complex dynamic systems and bringing automation to the deployment and the monitoring and the analysis of these systems. I like to think about AI ops as part of the full observability picture that takes this analyzed data, this root cause analysis and turns it into action. So, for example, you might have an application with a known, say, memory leak that might today take DevOps engineer logging into the platform into the environment, restarting the application or service. With AI ops, the platform itself would recognize the leak.
It could signal for restart and automatically take care of this without the engineering teams getting involved. This frees them up to continue on their important task. That's just a simple example. This could actually get more complex.
Agencies and educational institutions can greatly benefit from AI ops. So, staff augmentation, quicker problem resolution, creation of more robust and secure systems. Those are just a few of the arguments for an AI ops solution. So, next, let me take it up a level.
How do these solutions dovetail with the need for agencies and institutions to really strengthen their overall zero trust and cybersecurity strategies? This is a really timely question, especially with all of the hacks we've seen in the news and the ones we haven't seen. I've always been a very strong believer that observability and security, so desk ops go hand in hand. Security professionals create more information that really can help them understand more completely a security threat or vulnerability.
But this has to be valuable information since they're already bombarded with tons of alerts. So, having deep, complete, granular baseline data on all the operations of a system is extremely useful when looking for evidence of a system that's been compromised. But the question here, you asked us really about zero trust. I don't want anyone to believe that an observability platform is going to solve all of your issues around zero trust.
But there are many capabilities found in the best of class observability platforms that will assist in your zero trust strategy. And one of the key activities that we see outlined by NIS, that's the National Institute of Standards and Technology, or CISA, the Cybersecurity and Infrastructure Security Agency out of DHS. When creating a zero trust strategy, introducing network segmentation into the network is one of the first things you should look at. And this key step of segmenting the network is really fully understanding your application and all its dependencies.
So, true observability platforms are going to excel in creating this application dependency map because we need that for AI analysis. Now, these maps are extremely valuable when defining your network segmentation plans. Other elements like real user observability, the ability to characterize and even visualize what users might be doing on the system, kind of watching, not just from a performance standpoint, but also from an evidence standpoint, if something nefarious is going on. All of these are other important capabilities that a best of class observability platform is going to bring you.
I appreciate your laying that out for us. Lastly, can you share examples of where you've seen agencies or institutions benefit from the kind of solution that Dynatrace offers? I'll give you a couple. One success that I'd like to highlight is the state of Minnesota, and especially their economic development and unemployment insurance application, I think that was the proper name for it.
This application experienced a 10 fold surge in traffic due to COVID. Now, pre-COVID, this application was averaging 638,000 visits per month. When COVID kicked up, that peak to 6.6 million during the pandemic last year, averaging out after that peak to about 2 million per month throughout 2021. Now, due to the critical nature of these services, as you can imagine, the state faced really significant pressure to ensure that there was uninterrupted access for the citizens, but limited resources and the dependency on manual processes meant that the Minnesota IT team was spending hours in war rooms, kind of chasing down problems, which made it difficult to optimize the applications.
Realizing they needed a better approach, they engaged Dynatrace, and we helped them gain control of those apps to bring stability to the performance of those apps. We quickly partnered with the state, and in a little time, we had the applications instrumented providing complete visibility into the app so they could actually optimize the performance of that app, and we were even able to introduce automation into some of these more manual tasks, so that was a great partnership that we still have with them. Now, the second example here is, I'm just going to say it's a law enforcement agency, but I won't mention the name. And this was another instance where we had an agency that had a very critical application.
That application needed to be available 24-7, and to that end, they actually had multiple shifts. They had three shifts that would maintain monitoring over the system if there were problems here once it worked with Engage and tried to solve the issue. As you can imagine, three shifts running like the third shift, two o'clock in the morning system problem. It might take an hour or two to fix the problem, because once they diagnosed what the issue was, if they had to log into the system, they had enhancements or changes.
They actually had to get someone on the phone from the DevOps team who had the ability to log in and do that. What Dynatrace was able to do is take a lot of those manual tasks, things like restarting a process because there was a memory leak. We were able to take those manual processes, automate the root cause analysis, what people were sitting there actually doing manually. We were able to automate the remediation, and they were able to eliminate that third shift.
Now, elimination did mean that they got rid of employees by any stretch of the imagination. They were able to now retask those employees to more valuable tasks during the day, and then let Dynatrace almost act as the third shift, as a force augmentation of that department. So those are just a few examples of how Dynatrace has been leveraged by state and local agencies. Appreciate your illustrating how that worked.
I'm afraid we're going to have to leave it there for today, though. We'll thank you so much for joining us to talk about leveraging AI ops and observability platforms to help agencies and institutions improve performance and get to the cloud more effectively. Thank you. I'd great being here.
And we'd like to also thank Dynatrace for underwriting today's episode. Look for more of our IT modernization series on Statesgroup.com and edscoop.com and our Statesgroup and Edscoop podcast channels wherever you get your podcasts. This is why I cash. Thanks for tuning in.