AI in Splunk AppDynamics

Hello everyone, so my colleagues Jared, Isaac and I will be taking you through this introductory journey of the newly introduced AI capability.00:11Is in AppDynamics, namely the MCP server. If you are not aware of what MCP is or how AppDynamics will benefit from it, don't worry, we will try to demystify some of it in today's session. So Garrett and Isaac if you want to introduce yourself.00:20Yeah, I'll go 1st. My name is Jared mcnail. I work in customer success, specifically all around observable day, been with Cisco coming up eight years. Over to you, Isaac. Hi everyone. My name is Isaac Laurastett. I'm a colleague of Gerard and Tavish also.00:37Customer success in observability.00:57I've been with Cisco for five years now and I'll be in the background responding to your questions throughout the presentation. Thank you. Thanks. Hi everyone, so I'm I'm Tabesh I'll be doing the initial part of this demo today. I'm also in the same role as Jeddard and Isaac.01:01And I have been here with Cisco/Splunk/app dynamics for the last six six years now right and and I'm handling the observability portfolio just like including AppDynamics as well as Splunk observability now.01:21Okay, so during the session, we will really appreciate any questions that you have. Please feel free to to ask anything that any doubts that you might have. The best way to ask your questions would be to use the question and answer panel, right? Just click on the three dots on the bottom right of your screen to access Q and A.01:40Right. You can also raise your hand to ask the question, but I'll be honest posting it on Q and A is definitely a better option since, that would mean that the question, your question is registered and somebody gives you a proper answer like our colleague, Isaac will be there in the background answering all your questions, right? And.02:01That will not also interrupt the flow of the session. So really appreciate if you can put your questions on Q and A, but just keep in mind that the the that you know, you can also raise your hands in in in situations where something is very urgent.02:20Okay, so the agenda for today is we are going to talk about AppDynamics MCP. 1st of all, we'll just quickly cover what is MCP in general, and then we'll be talking about AppDynamics MCP. Then why did AppDynamics roll out the MCP server, right? What was the reason or the rationale for that? We'll be discussing about that.02:38That they'll be we'll be talking about the how like the architecture and the design of MCP server of appdynamics. And finally, my colleague Jared will do a live demo to show you all the capabilities.02:56In live action. Okay, so before we start, 1st question, alright if if for those of you who have joined later or who have not yet scanned the scanner or the QR code is on the left hand side.03:10Right, please scan it or open it on your browser and we are eager to hear what are your experiences with MCP. So we'll just give it almost a minute to see what kind of.03:25Understanding you have? We've got a lot of people using the Slido, which is great. Thank you. I'll just give it another minute or so. Yep.03:45But it's the reaction to the question is positive and that hopefully the next couple of slides will help explain what AMCP is and from a general point of view and then specifically we'll go into app the Splunk AppDynamics MCP as well.04:05We got a great engagement here Tavish. Yeah yeah so there are a lot many more attendees, so we'll, we would appreciate if everyone can answer this question.04:24Very good, NO, yeah, we can move on to the next one as well. People are still joining, they can answer it. Yeah. Okay, so the, so what we understand from this is that the overwhelming majority of people have not used MCP before and that's completely ok. The the whole.04:44Point of this session is to understand try to trying to understand and trying to see what AppDynamics has provided in terms of MCP integration. Ok, so I'll I'll just go to the next question now.05:04Okay, so I would assume that, you know, since the overwhelming majority of people are not aware of MCP and this is a new feature in AppDynamics, probably not many people would, would be knowing about this one.05:21Right, by the way, this is one of the relatively newer features, new offering in AppDynamics, so I do not expect many people to even know that the new feature.05:36Is is is already live on on AppDynamics.05:56Yeah, that's a really good point Tavish, and it's probably reflective in that the response to this question that this specific feature, the AppDynamics MCP was GA.06:02At the start of the month. So what's very, very new. Yeah. So shall we go ahead.06:17Yeah, please. Yeah ok so at the very start, let's let's talk about you know the overall concept, right? Before digging into or diving deep into what AppDynamics provide in terms of MCP. What is MCP here, right? MCP as many of you already know, even if.06:33You have a basic introduction is model context protocol. It's an open standard that has been developed by anthropic.06:53And it lets artificial intelligence applications connect securely to external data sources, tools and workflows, right? So think of it as, the USB port for AI, right? Just connect your AI to those external systems from which you want to extract or retrieve the information as simple as that.07:00Now, in terms of the architecture, this is how the MCP works, right? As you can see, you have any AI application, any LLM client is there, right? And that connects to the MCP server, and we'll come to this in a moment that you know what an MCP server is, and.07:19Yeah, so the MCP server in turn uses or leverages the connection mechanism which can be the REST API or, you know, JDBC drivers, it can be anything, and then it connects with the external systems via those REST APIs, e.g., and then the response is returned to the, to, to the.07:39The application that has requested and the application processes and tries to make sense of that data, right? So this is a very simple flow of the MCP client and the LLM to MCP connection.07:59Now what is AppDynamics MCP server? Right, so before we go into this one, let's let's just quickly let let me take a moment to start to to mention that MCP is only the newest addition to appd's AI based capabilities, right?08:15I would like to remind that our dynamics already has many AI based features that you have been using for some time. Anomaly detection, right? Dynamic baselines.08:33Then automatic root cause analysis, right? All these are AI based features that have been available in AppDynamics for quite some time, right? And now we are aware that many of our customers of our attendees today are, have on premises deployed. We traditionally didn't have access to the AI capability.08:42It's.09:02As well as the SaaS offering. But the good news is that the new version of AppDynamics on prem, which is called virtual appliance or VA insured, has access to all those AI capabilities, right? Now if you are interested in using virtual appliance, if you are an on prem customer and you are not currently using V.09:03But you are interested in using it, please reach out to your customer success or accounts team if you want to know more about the the other AI feature.09:23Okay, so what is the feature here, right? Now, appd MCP server is an integration layer that enables AI agents to interact with AppDynamics data and tools in a structured, secure, and permission aware manner. So it's not very different from the overall definition of MCP that we saw on the 1st slide.09:33Right, right. So basically MCP in AppDynamics will act as a bridge, allowing AI applications, whatever AI applications you're using to query and invoke AppDynamics capabilities through a standard standardized MCP tools.09:53Right? If you are already using MCP form factor in your organization, the good news is that this is less of a learning curve, right? Because you know how to use the MCP. You just need to connect AppDynamics to your MCP platform, right? And you can immediately start leveraging the AppDynamics MCP features.10:09Okay, so, this is you know very high level view of the MCP server, right? You have your AI assistance, right? And, you know, it can be anything that you are currently using. It can be a cloud based tool, it can be chat GPT, right? Or the codecs environment.10:29It can be GitHub copilot, anything that you are using currently, you can just plug it into the app DMCP server, right? And the APPDMCP server will do the heavy lifting of connecting the observability data or retrieving the observability data for you and making sense of that data for your AI platform.10:48Right, ok, so, now here we you can see that all these capabilities and tools are mentioned here, right? But, we are talking not only about the features or the components of appd MCP server, but we are also.11:08Highlighting the possible advantages of this approach in this way, right? Using these provided tools and prompts that are part of the MCP integration, it can do the heavy lifting for you. Looking at different parts of the application and establishing the core.11:28Relation.11:48Observing things that probably a human eye may have missed on the UI. So e.g., when you're using one of these guided flows here, right, the prompts, the prompt will help you to surface any kind of issues or troubleshooting hotspots that are there in your application, right? Without you having.11:49To manually solve through this on the, on, on the UI.12:09Right, and then these these are the things that will become clear once Jared does the demo for you, right? Where you will be you would be able to see that how these prompts and the tools that are available at your disposal can be used in order to get the information or you know insights into into the application.12:15So, now this tool, as you can see, this is GA now, right? Appdynamics MCP server is now GA. It's generally available since 1 September, right? For those of you who have been part of the alpha program before the GA happened, you will be happy to know that a lot of capabilities, tools.12:37Prompts resources have been added based on your feedback. In addition many new troubleshooting domains, like I think during the alpha program we were doing the APM troubleshooting domain only, but a lot of new troubleshooting domains like EUM and database monitoring, cluster monitoring, et cetera, have been added.12:57You will be able to see few of these troubleshooting domains in the live demo that we are going to have later during the session. Now the question is, why did AppDynamics decide to roll out the MCP feature, right? What was missing in AppDynamics, the UI part right or or for that matter.13:17The REST APIs that are available.13:37Well, the answer is that SRE teams today are evolving very fast, right? It's an AI 1st world. Everyone is trying to make sense of everything that they have in terms of the tools and the capabilities.13:41And how to utilize them using their AI capabilities or using their AI tools, right? So e.g. moving from detecting symptoms to diagnosing diagnosing root causes, how do you troubleshoot an application when you are on the UI?13:58Right? You look at the symptoms or anomalies and then dig more to arrive at the root cause, right? This is what we do when we are on the AppDynamics controller or the UI. That's the logical flow of troubleshooting and there is nothing wrong in doing it that way, right? You go, you try to find out the symptoms.14:18And then finally you arrive at the root cause. But what if there is a better way? What is the tool that you're using can give you the root causes straight away and then explain the rationale behind the decision. This is what the MCP approach can potentially do for you, right? It can not only give you all the data that you are able to see.14:36On the UI, right? But also give you a root cause much more quickly than you would on on a UI when you are doing things on your own.14:57Right. Now, who has access to the features? Customers who have licensed for appdynamics, basically all AppDynamics customers, whether SaaS or on prem, already have access to these features. Now the release timeline is 20 six.seven, right? Which means that you know any any any.15:09Controller version that is 20 six.seven onwards would be able to utilize the features, but I guess I think even earlier versions can utilize that. Is that correct?15:29That is correct habish. Yeah, and I think that's very important to point out not to give the wrong information that the way of all the the functionality of the AppDynamics MCP server, 20 six.seven would be your best release.15:40However, if there's customers here in the audience today who have earlier versions, you know, 20 six.one e.g. 20 five.ten, they can still avail of this, of the AppDynamics MCP, but they may not have access to all the prompts and the.15:56The APIs and the features which a 26 six.seven release can show basically long story short, it's not a showstopper, if you're not on 20 six.seven, you can still use the feature if it's if you're on an earlier release of a controller.16:16That's good to know. Thanks Jerry. Thank you. No problem. Yeah, so now the question is that how can you start using the feature? The answer is that the the build for MCP that is already available on the Splunk AppDynamics docker hub portal, right?16:33And we do have plans to introduce it to the download portal as well, right? So very soon this will be available on the download portal, but for now you can download it from the Splunk AppDynamics docker hub.16:52Yeah, ok, so, now this is a you know some further background to why we we introduced the feature, right? I've already talked about this one, but then you know the typical problem that come with observability domain, right? Data silos, insights are tracked across disparate tools, right?17:06Manual correlation, right? You have to go between multiple UIs, multiple dashboards in order to make sense of what you see, right? And high toil of the 1st 30 60 min of an incident when you have to manual manually gather the data and that indirectly inflates or increases.17:26MTTR for finding a resolution, right? So all these things, they will, a tool like MCP you know that helps to integrate AppDynamics with your AI tooling will help you to take care of all these issues with a typical.17:46You know, monitoring on the on an observability UI, right? So simplified onboarding like for those people who are not probably who are not very aware of the nuances of how AppDynamics shows everything, right? MCP basically has.18:07The potential to democratize access to appd, even those teams who are not very comfortable with the appd UI, they would be able to make sense of the data that is being thrown at them by helping the MCP to sift that data and to give them clear insights about that data.18:27Right. So, your AI agents will be able to uncover additional visibilities quickly from the MCB server. Okay, so, from here, let me let me just go into the architecture of how this has been done.18:47Designed. Now this is a very straightforward architecture if you try to understand the flow here, right? This is the MCP client, right? At the top, at the very top, this might be your, maybe your cloud desktop. This might be cursor, this might be something like GitHub codecs.19:07Or a Microsoft copilot, right? Anything, any, any of the LLM clients that you're using, you can connect it to the MCP server. Now this is where the MCP server, AppDynamics MCP server is, right? Now this is, as I said that, you know, you can download it from the docker hub, right? And there is a.19:27The configuration that needs to be done. Once that is done, you will be able to connect your MCP client to the MCP server here. And MCP server in turn will be talking to the AppDynamics UI or sorry, the AppDynamics controller, not the UI, right? How does it talk to the AppDynamics server? It talks to the REST APIs. Now there can be.19:47Take several mechanisms in which the MCP server talks to the backend systems.20:07We have chosen the REST API because we already have very rich REST APIs available, right? And these are the REST APIs that are retrieving the data from AppDynamics and sending it to the MCP server and MCP server then shares the these details or that correlated data to your MCP client.20:13This is the on prem one, right? This is the on prem set up here, right? This is similar to what we have seen here on the one, so there is not much of a difference, right? As you can see, this is the fast NCP server, and then it talks to the on prem AppDynamics controller. Now, one thing I would like to highlight here, which is very important is.20:34Many of the AI capabilities that are built into AppDynamics may not be available on on the on prem setup, but then MCP is, is, is an exception. MCP is available on both the versions of on prem, both the classic on prem, which is the old on prem.20:55As well as the new virtual appliance. So if you are already on the classic on prem, which I guess most of the customers today, most of the on prem customers today are on, you don't need to worry MCP client will work on the classic on prem as well. But if you want to take advantage of the other AI.21:14Capabilities only then you have to go to the virtual appliance version of on prem.21:34Okay, nice. So, now is the time for the demo and I will hand it over to my colleague Jared to take you through a journey of how this works on a real environment. Fantastic. Thank you very much Tavish.21:39That was really informative and hopefully the audience got the theory behind what we're, why the why and the what behind the the decision by product and engineering to introduce that that's MCP server 1st around this whole AI 1st world, so I'm gonna stop talking and share my screen.21:57And then we'll get down into shorter, shown you in practice.22:17Oh, let me share my screen here. Okay, and 1st of all, I want to go and basically show you a controller, new vehicle right away to show me that. Basically, I have my I have the MCP server installed.22:22On my laptop on and I have docker desktop running, so it's running within the docker.22:42The controller we're gonna talk to basically Interrogate is a demo controller, an internal one. So this is the one here we're gonna talk to, and just want to run some commands on my API client and then show you the controller just to show that it's not what they call an AI word illucinating and give you.22:49Wrong information e.g., so it's quite re, you know, it's very reliable the MCP server. I've played around with it myself. I'm actually really impressed by it. It's.23:09If it's really about this way of how you're looking to change how you work and if internally within Cisco we're using AI quite a lot just from an operational role for, for ourselves.23:19And I'm just thinking for our customers that that use maybe on that journey, you may not have started it, but it's just to make you aware that if you do, we're there to support you by this feature. So, I want to bring up, let me see here some I don't want to show you that.23:34Yeah, this is the client here, but I want to show you, documentation as well. Here is the tools and prompts and why you better saying, right? Why am I showing you this? Basically, this is what I'm gonna run. Not all of them because we haven't got the time, but Isaac I'll grab this link and put it in the chat.23:53And it's in the docs as well. These are all the different prompts available with this current release of the MCP server. We are doing a lot of stuff post GA as well, specifically around e.g. the the cattle, the maturity assessment tool, which I know a lot of customers are aware of. You actually can run this.24:13That we plans to run this via your AI, AI client. So this is a link here just around the tools and prompts which are available in this release. Again, we're looking to expand that in future releases. So.24:33Let me just bring this up the full screen. Can everyone see that? Are we all good? Give me a yes. Yes, yeah. Brilliant. I always have to do that because sometimes I've talked and people will be panging going, I can't see your screen. So I've learned the hard way.24:49So this is codecs. This is the the AI, the AI client of, of my that I've chosen to use for this demo. It's very clearly made, it can be anymore, it can be any flavor, it can be cloud, you know, whatever, you know, it's totally up to you as I'm just.25:10For this demo. So the 1st thing I'm gonna do is just type in codex's MCP list and you can see here it's franken. No, I'm starting very small here folks.25:31Basically just to show you that on your client, you have, you can configure certain MCP configuration. As you can see here, the current MCP configuration has unchanged I have this set up from yesterday and but you can see the appdynamics one. So that's enabled, it's got bear token.25:41As Tabush showed in the HLD slides.26:01That you have to have a token to access, to talk to the controller so you have that security as well. So that's very basic. That's just we're talking, we're not we're just making sure that it's configured and it's and it's set up to actually talk to the controller of choice.26:04Now, I'm gonna like just put in something like here and this I was gonna cut and paste here to save time, but I just want to show you like can you access the the app dynamics MCP server tools, see what this is here. So we're doing we're sort of what we're doing here.26:23It's just doing probably checks.26:43As Tavish mentioned, there's some configuration, not a lot, but there's some configuration to actually set up your client to talk to the MCP server to talk to your specific controller. And I think a really good way to, to think of this overarching AI architecture is that.26:46I the MCP server is a bridge.27:06From your AI client, so that AI 1st approach to talk to your controller, the MC server works as a bridge to be able to communicate. So you can see here yes, the AppDynamics MC MCP server is enabled and accessible with bare token authentication. So there is there is specific.27:09Application which you can set up and control so not everyone can access it. So you have that layer of security. And then I can say, ok, that's good. How many tools are available, say.27:28And by the appdynamics. So you were told you were just using, you know, basic English here just to talk to it, you know, that's that's what I love about it is, you know, it's not, you're not following a strict.27:43Sort of syntax they'll be able to get information from the controller itself. So again, this is just a check. We haven't really got under the meat here and we don't worry we'll get there very soon is there's a hundred and five tools available through the AppDynamics MCP server.27:59To allude to the point Tavish mentioned earlier, there was an alpha program around this feature and some customers were involved. That was very limited. There was only approximate, I think there was 33, if I recall correctly, so we widened that out to cover a lot more domains.28:18So you you know those who are part of the alpha program will see a big change in the GA version. So let's let's get going into it and I think I've shown you enough around the, you know, the, the prechecks and just to show you the talking and what's available. So I wanna go I wanna ask what are the current.28:37PM applications instrumented instrumented on the SE lab controller. Remember I showed you that previously that's a controller we're talking to today and we're using for this demo and I want to see.28:57Just as you know you might, you might go on in the morning, you might or someone to go someone maybe doing a review. You've doing an audit e.g. of their controller, and you just wanna know right how many actual applications have we own that controller and configured? And as you can see here, it's returned a 958.29:15Configure application and this is like so it's basically an inventory of what what you have and let me if I can show you that. You can see here this is the ASC lab. You can see applications 958. So that's correct, basically. So that's why I wanted to show you.29:35That so and you, there's still different applications you can down get down in the meat on that. You can also.29:55Maybe look at, and this is away from the alpha program, this is APM, so what if I went list, you know, the the BROWSER applications.30:05With real user traffic and response time, see what comes back on that sort of curiosity and again these are all single prompts. I shared with you that the I think Isaac shared with you the length there.30:24Off the tolls and prompts, there's a huge number. Again, as I mentioned before, we are looking to expand that as well post GA, but this is again, this is probably just getting a feel of right what's configured on my controller. Sometimes, you know, a large controller controllers can have a huge amount of services.30:41Either from the digital experience monitoring, your browser, your mobile, joining to APM, you know, your database, you know, cluster agent, machine agent, all that's covered with NSTA release. So you can see here's.31:01Oh, this is taking a bit longer. I think it's it's fair to point out things don't happen right away. It takes a while to unless the EAM is telling you what it's doing. It's identifying the EAM applications, but it's identifying with real user traffic, so he wants to look at, you know, which actual applications have real user traffic on this demo controller.31:19A lot won't because it's a demo and we, you know, put, you know, certain traffic through certain applications and maybe non free dollars for days, weeks, you know, depending on the setup and for that on the controller itself. So arguably it's not the best for a real life, you know, controller, which.31:39Just Instrument that within your observ observable layer state, but here you can just try and report the response time metrics, it's going over a current aligned window. If you don't mention time, it usually defaults to one order.31:59You know, so it's probably gonna be less than the real user traffic and response time for the last hour, but you'll see that once it comes out over once it completes its processing and investigation and then you'll see how long the timeline, but you can specify a timeline accordingly in your question.32:14And, like, e.g., I'm gonna while this is running too, I'm just gonna as you can see, it's given me the time, it's telling me what it's doing is querying the metrics, I think there was a question in the chat around does it, you know, look into metrics in advance? The answer to that is yes, it does, you know, depending on the question.32:34And what you're and maybe the use case you're trying to get retrieve from the controller itself. From a SaaS perspective.32:54We also, it's only read only. We're not changing anything here just to make that clear, and I think that's important around change management. We are literally just retrieving information from the controller in question and then, you know.33:02Displaying it on your on your client. So you can see here that talk about 2 min 40 s. I'm gonna be honest here. Full transparency, it depends on the question. It can be less than that, it can be longer, but, you know, it's usually quite quick, you know, as again depending on, you know, what, what you're asking the controller to do. It's.33:19Underpinned by APIs, so I think this tab is explained and that I showed on previous slides, you can see it's all API, so I'll just reading that and calling them and displaying the information. So you can see here, as I mentioned, it's all it's used in the last one here. If you don't specify.33:40That time defaults to 1 h, ok? So I guess it's important. Obviously, if you go the last 6 h or the last 24, it won't take longer to process that because obviously there's more information there and there's more metrics to investigate before I can respond to your question.34:00You can see here the app, app key page request and user response time, you know, and currently has the highest response. Tell metrics will add it to 1 min conference, so there's a bit of a con, you know, a conclusion as well to the question what you're asking. What I want to show you here and just how.34:19I'm just let I want to go and do maybe.34:39You know, look at a specific application rather than just go controller wide which we've done here. I want to go sort of maybe deep deep down into a specific question, a specific application. Sorry, so like troubleshoot slow response.34:43For whoops it is a for a AD e commerce, ok? So that could be something if you're as, No I'm trying to think of right you maybe watching this now and going right, how would this.35:04Make a positive impact to my role and whatever role, you know, you guys do, it's all wide and varied, but if you're like operations or you're looking after a specific service or application, you could come in the morning and go right, you may have your AI client, you know, turning within your enterprise and go, ok, I want to have a look at the, you know, I want to trouble.35:21You should slowly sponsor AD commerce, you know, and see if there's any, you know, the behavior of it. Now you can see here it's it's trouble, it's telling you what it's doing. It's checking over the last 24 h.35:40Because to troubleshoot slow response, it's not a single query. This is a prompt, so we have multiple prompts available on this GA release, which basically calls multiple single queries, APIs to respond to that question. So if you came in and if your manager was telling you, look I want you to.35:51Look out through the service, it's you know, it's.36:11We're having a lot of negative feedback from our end users. You could come in the morning go right I want to, you know, troubleshoot response for 80 commerce for the last 6 h, if you didn't want to do the last 24 h or for the last 12 h, and then it gives you know quite a quick response and quite a, you know, high.36:15Level overview of, of the health of that. That's even before you jump into the controller.36:34So this is this AI 1st and this is what we're seeing with a lot of our customers is that they are moving some quicker than others to this AI 1st approach, and this is what we can do with the MCP server, if you wish, is that, you know, we have the capability to use your specific AI clients.36:39To talk to the controller by, as I mentioned, that bridge.36:59No, it's often that we've seen with a lot of even with the alpha like you go, yeah, we have to install this, you know, we also have a capability in alpha at the moment around AI assistant. I think it's sort of, it's a parallel with MCP.37:03And the reason we introduced it's back in with the controller as well. It's not GA, it's alpha. Why am I imagining that? Because it depends on how you want to approach and how you want to, you know, take control of how you're looking to deploy that, those type of AI processes will.37:20And your enterprise.37:40The MCP, in my pa in my opinion gives you more freedom to do that. You can play around with it, you can modify it and you can set it up with whatever AI client you wish. The AI assistant, which is as I say, currently offer, that's baked in with the controller, so you you haven't got that more freedom or flexible.37:42Around how you want to use it, but it's it's plans that gets year'd over the coming months.38:03But as you can see here this has taken a bit longer. Why? Because it's a longer timeframe. As I mentioned, it's not 1 h, it's the last 24 h, and you can see here it's it's like it's it's not just like what I like you know sometimes when the dread of the spinning way of, you know, doom when you're trying.38:08Trying to do yourself on your Mac or whatever your laptop of choice is, but here it's trying to explain the ok we've used app D integration, we're trying to get the context here of.38:28Of the health of that specific application in this in this case it's AD commerce and it's just giving you, and it's checking over 24 h, so you it's it has a lot to gather before it can respond accordingly, which is good you know like you know you could literally type this in and.38:42All right I'm gonna go and grab a coffee and it should come back in all within a couple of minutes then. But to see give a time and then we'll we'll also show all our, I'll show you the all our potential we, what we can do once we get this output responded back to us.39:02There we go, there we go. So how long does that take? Just over four months, so you sure I'm not trying to be I did say it though it's not right away, but it depends on the question you're asking. So let's have a look at the output here. So it's going it's look at the AD commerce response is concentrated on the checkout business transaction rather.39:19Being uniform across an application and I like this table. It's giving you an application overall assessment, checkout BT, it's called that specific BT, right? So right away.39:39You come back after grabbing your coffee and you sit down and go, ok, I might want to look into this, check my BT that this is the the power of it, you know, you can ask it in plain English and they can come back around, you know, maybe red flags you want to investigate further. A good way to start, you know, the morning.39:51Especially if you've got a responsibility on a specific servers or application to make very, very powerful, and you haven't even went onto the controller here. This is literally all in the on your AI client. It's called, it's looking at tax calculation. Now this is specific to let this application and service, it's important to call out, but it's looking at the class and.40:10Effort and how long that's taken up.40:30Again, another, sorry, red flag that's for you to investigate, looking at the database look up for a chart as well, how long that's taken as intermediate database related latency mode. So it's giving you what the areas it's looking at spring security and you know they have.40:33From the last 24 and then its own assessment.40:53And then it gives you sort of conclusion here at the bottom, the main diagnosis is too long tailed latency modes. Again, that tax calculation, so that's that tax calculate, that's basically the class and method, so maybe that could be improved and we can look in that further Payment of fulfillment says also show latency and.40:56What currently happens doesn't, you know, provide the primary checkout.41:16Bottleneck. So basically it's really, you know, impactful. It's giving me a conclusion on where, you know, you could start investigating a versus certain issue around your service and application. Now, if I wanted to do maybe, well.41:20What is the cause of the checkout of the checkout BT performance performance?41:37I can just put that also check IPD. So I'm trying to I'm sort of laser from what I put it to me, I'm trying to understand the primary cause of the per check IP, so it's came back quite quickly on that. So you can see here's.41:54Talking that sort of doubling down on the result that profile that's shown earlier and you can Sorry, understand where you could maybe look into potential problems and around your investigation, so.42:10From a day to day roles and right, how can this help me? You know, how can this make a real difference? Basically, like you're, you're, you're looking to prove your MTTA, do you know what I mean time to identification, if you can just type this in and you're gonna have and as a result, your MTTR, you can look at RCA.42:27So all these specific use cases are rules you may have to, you know, meet in your day to day job. This.42:48Functionality of this feature can definitely improve that, in my opinion, and from experience and feedback we got from the alpha program, we've got that from our, you know, large enterprise clients as well that they do see the benefit of this. So.42:59What else we could do is like, I could look at, you know, we got the checkboid, you know, maybe what is, you know, what are the, the snapshots?43:16To check like PT performance because that's what it's doing. The APIs are not modifying anything. I think that's very, very important the you know that's a point I made earlier, but I want to reiterate, it actually shows you, you know.43:36It's just basically retrieving information. But see you can see here now I just wanted to go right ok how did you keep come to this conclusion, and it's called out here you know the sort of the represent the term represent check check out snapshots again from the last 24 h.43:57And and team were analyzed in detail and has shown you here and you can like you know copy that and have a look. So again, it's showing you the ones that analyzed in detail and then goes, yeah, and this is why we come to a conclu conclusion. So you're sort of questioning it because personally, I question.44:13Everything AI says, that's just me. I use it as a steer and I think that's important, but that's just my general attitude towards AI, but yeah, it can.44:33You can double down on it and go right, ok, how do you find this as a direct me towards something, you know, so I think that's important, but that's just a way high, I work I work with AI but again it's just showing you the list of snapshots, you can jump onto the controller, find the snapshots accordingly and you know.44:44Confirm that again making sure it's not illusinating and all it's not giving any, you know, wrong information.45:04Oh, just I've been talking a lot and like this is only a real snapshot of where we're at. I need to stop now because I'm just conscious of the time got 10 min here. There is a lot more here folks, a lot more, and that's why I wanted to share the link around the tools and prompts. There's, you know.45:11I can get like.45:31Get health rules, oh, you know, for again, oops, my typing's not the best. Let me see here.45:33So I can just go right. I mean it's very specific just for AD commerce, but that's probably reflective and I'll retrieve the current health rules configured. Again, it's a more of an audit. That's good, but I think the more real powerful is when you start asking, you know, the troubleshooting and look at slow responses and then look at BT performance.45:52You know accordingly, you can just instead rather than just getting the help flows, you could go right show me the help flows which have been activated in the last 1 h e.g., you know, you can see here it's just less than ones which are configured when is enabled and one what's disabled and the backend again really good for.46:12For an audit purpose, but if you go show me the health rules activated in the last 1 h. So I'm using time here, I never used time previously in the question.46:33No, I didn't mention ADA commerce, but still cause we're in this conversation that just picks it up by default, which is good. Rather than just running through for the last hour and it's just gonna again this is a really good way if you go right I just want to type this in because I've just arrived and 1st thing in the morning 09:00 and go, ok, I wanna have a.46:51Look at say any health rules violated and you can just talk to it and then you can jump onto the controller to do deeper dive, you know, accordingly.47:11Right, I'm gonna There we go, and it just shows you always red criticality commerce, those controls are activation part are NO longer active. So it's actually telling me ones which were activated, but are NO longer active ones which are on on ongoing, and it gives you the severity, et cetera, and it gives you the effected the entity as well.47:19Okay, so folks, that was a very quick demo. Again, it's literally only touching the surface of the capability of this feature. The point today was to highlight this to make you aware and really maybe to grab your interest and, you know, and if you want to learn more, please reach out. As you can see here.47:41There, these are a couple of questions we asked near the end just to gather, you know, feedback from the audience if there's any specific topics you would like to hear more about in the future. We have another session coming up next month.48:01It's the product update session. Usually very well attended because it covers all different types of updates made to the controller over, you know, the last year. So maybe some features you might may not be aware of, so we touch on everything and that covers SAS.48:14OVA and on prem. So it's not concentrating on one specific architecture as well, but this is good, you know, to get this API usage and adoption. Yeah, we can, it's all AIA, you know, the API usage and adoption, yeah, we can definitely look at, you know, double looking into that. So we're getting a good.48:32Response as well.48:51I'm gonna go to the next one, so I'm just conscious of time. So here is you go under Slido, you can type in any topics you would like to more hear more about. If the previous question didn't have areas which you maybe interested in, please just type it in here and we'll take it on board when we look at plan and future.48:53Conversations as well. This is great, you know, what's the point in all springing something up which doesn't get you guys here to hear the everyday users of, of, of our platforms and all engaged, so we love to hear from you guys. How does this ties in with Splunk? Yep. Good question. Yep.49:13Yeah, yeah, this is very appdynamics, but we have this falls into the whole Splunk observer of all the portfolio. I think that's important to call out, so that's a great question. Call Leo, yeah, again, great one. Get a lot of feedback from after our.com last week.49:30Layer.49:53Longer and troubleshooting database performance. Cool. Yep. That's another good one. We can definitely do that open telemetry, yeah, and that's behind the whole, that's a future hotel and that's what we see around, you know, the whole Splunk observer well their portfolio as well. Oh, this is really good.49:54Thank you all for the engagement and how I'm thinking Spark ITSA. Okay, yeah. Yeah, I think it's just looks like that's the big one. That means a lot of people take that in and they're all putting.50:14That's the one that seems to be getting a lot of feedback, so that's interesting, so that's something we can definitely look at. I think these sessions have been very Splunk appdynamics. I think, yeah, going forward, we're gonna cover all the Splunk observability portfolio as well. So yeah.50:32That that's how you well widen that scope out, but this is great. Thank you very much for that. I wanted to call out just a couple more slides, Cisco U, that's our Cisco university. It's just basically a repository where you can, where it offers self paced.50:49And instructor led courses, so it gives you know a deeper understanding of the capabilities, that's just from a wider Cisco and also observability perspective. So that's, you know, if you're not aware of it, please go to you.cisco.com and you'll see a lot of options there around training.51:06I love calling this site because I just love this channel. It's it's a YouTube channel. It's at Cisco observability, lot of content there covering numerous areas and it's.51:27It's all just like their long videos or some of them are 45 min long, but there's some or two or 3 min, which is what I love, it's very ingestable, it covered various use cases. So if you're not aware of it, please, you know, check it out and there maybe some, it might help you round your self paced training and understand.51:42On an author platform as well.52:02And then finally, it's just the resources of what we called out. That's the length of the YouTube channel there and documentation for Smart AppDynamics and the whole observability portfolio. And then at the bottom, specific to today's session around the Smoke AppDynamics MCP server, you've got the length.52:06To the documentation and also the link to the tools and prompts which we I showed you earlier.52:25So, yeah, please have a look, play around with it. If you have any questions, please rea reach out to your customer success representative or your account manager like maybe if you want to have, you know, follow up sessions or deeper dives cause conscious.52:31Again, this is more of a.52:51This is what we've just released as Tavish mentioned a couple of weeks ago, and it's just literally been out the door and we just want you guys to get engaged with it and start using it and see what you think and your feedback will help direct, you know, the future engineering investment into the specific feature. But as I mentioned, there.52:53There's a lot planned already for next month. There's a new release coming out next month.53:13And that's I'm actually quite excited about that because I mentioned before it's gonna touch around, you know, like the maturity assessment and the capability and prompts around that, which you can use via your AI client of choice. Okay just one one quick thing. So you, you might have noticed in today's.53:17Demo part right that Jared was running that it takes some time for you to to to experiment and to be able to get used to the new flow or the new way of working with the MCP in order to retrieve the information from appdynamics. But then over time, as you become more.53:37More comfortable, hopefully it should become easier for you to use the MCP interface confidently in order to pull the information that you want, right? So please, Isaac has already shared the link on the on the chat. Please feel free to download the MCP server. Experiment with it. If you have any questions, please.53:57Reach out to us and we would as Jared said, we would really want to hear your feedback on this one. And if you are having any teething issues, then we will be happy to resolve them for you. Yeah, number percent. A hundred percent, yeah.54:17Okay folks, we're right on the oil. We just about made it. So, thank you everyone for, attending. I think it's in Dharmesh, you've raised your hand. Do you have a question? Alright.54:36If you have a question, you can go off mute if you, NO problem. And remember, you can always reach out to us. So yeah, we'll be available. I Jared and Isaac will be available for any questions even after the session. Yeah, a hundred percent. So thank you all for taking the time out. We we do really appreciate it. We had a fantastic holiday.54:53Today, great numbers.55:13So we hope you find it really useful and it was, you know, a good use of your time cause we're conscious that, you know, everyone hasn't got a lot of time in today's world, so appreciate it. Thank you.55:15