.Net Foundation On .NET, 26 Apr 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Analyzing your applications with Application Insights

Description

Application Insights is powerful Application Performance Management (APM) service for developers building application hosted in Azure, on-premise or elsewhere.

Get your questions answered on the Microsoft Q&A for .NET - http://aka.ms/dotnet-qa
Learn .NET with free self-guided learning from Microsoft Learn: http://aka.ms/learndotnet

In this episode, we are joined by Michael Milirud (@micwebnet) who is a Product Manager on the Azure Monitoring team. He shows us how easy it is to get started with Application Insights with Visual Studio. He also discusses a few new features that you can make use of today such as Live Metrics, Load testing, and more.

01:05 - When can Application Insights be used?
02:41 - Exploring the ApplicationInsights.config file
03:56 - Enabling the Application Insights experience in Visual Studio
05:13 - Live metrics in the Azure portal
06:09 - Adaptive Monitoring
08:40 - Portal experience
12:10 - Performance monitoring
14:03 - What types of applications can make use of Applications Insights?
15:10 - User Analytics
18:15 - Application Analytics with Kusto

You can learn more about what Application Insights has to offer, how to use the Kusto query language and more using the links below.

What is Application Insights?: https://docs.microsoft.com/en-us/azure/application-insights/app-insights-overview?WT.mc_id=ondotnet-channel9-cephilli
Azure Application Insights REST API docs: https://dev.applicationinsights.io/
Kusto query language: https://docs.loganalytics.io/
Create a Free Account (Azure): https://aka.ms/azft-net

A

In this episode of the on.net net show, we're going to learn all about application insights from my friend, michael.

A

Welcome to another episode of the on .net show., my name is cecil phillip, and with me today, I have michael millard.. How are you doing michael?, I'm doing great.? Thank you for having me.

A

thank you.. Thank you so much for coming on,, man. so,. Why don't you tell us a little bit about what you do and what team are you on right now in microsoft?.

B

I'm a prog manager on azure monitoring, team.

B

Okay.- and there is a number of products we all put together for the azure monitoring experience., the one I'm working on right now is application. Insights.

B

okay., as application insights for me, is.

A

Probably one of the ones that I'm more familiar: with., yes. and so,. When you talk about application insights,, it's like application,, performance, and monitoring tooling, right?, like that's what it does?.

B

That's correct. so,. There are multiple layers. You could be monitoring when you run production workloads., you could be looking at infra,. You could be looking at the azure network, layer. sure.

A

Or you could be going all the way up.

B

The application layer where you see requests coming in, dependencies going out, exceptions happening in your code., got it. so,. I'm used to writing logs,.

A

I'm used to capturing telemetry, and some of these type of things. is application. Insights, a development time tool,.

A

Or is it a production time tool, or is it something I could use in both places?? That's the beautiful thing.. It's actually both.

B

You can use this both during interloop development when you write your code inside your id, and we have a rich integration story. There.- and you can also use this when you're monitoring production workloads,, whether you have access to your code or not., got you. got you. got you.

A

All right,, so why don't we take a look at the demo? You have here? sure. so,. What's this?.

B

So,, I wanted to show you something that is sort of very close to the dev experience.

B

This is an app service.. It's a demo, app service that we typically demo ai with..

B

It's called fabrikam fiber and I could click around this, and it's a basic ticketing system for employees.

B

Okay. right., some of these things will fail. like. If I try to do this,, it will fail, there's an exception.

B

but overall,. It's a very simple app- and I have this all here in visual studio. so,. It's just a regular project.

B

and one thing to know about application. Insights, we've not only worked to integrate this throughout the dev experiences,. The visual studio, experiences.

A

B

Which means,, if you create a new project,, you'll, see a toggle there at the creation path, saying on or off ai.. If you turn it on, it'll, just work., that's interesting. so,! As.

A

I'm creating my new application,. I can just flip the switch, and I can turn on application insights?. That is correct. and if you forgot that,.

B

You can always use the wizard to play it right, afterwards, or you could go to the portal and apply it from there as well. okay, great. so,, that even if.

A

I have a pre-existing application, when you're on azure, just inside of visual studio.

B

A

If I missed out the opportunity to get it on creation time,, I could always put it in a little bit. Later.

B

That's correct.

A

Cool, great., so in this case, there is.

B

Already application insights enabling this project, and there is application, insights, config, file., that's where all the wizardry of what we enable and pull into the project. Goes, and you're welcome to look at this afterwards..

B

There are a couple of handy links here that can guide you to various experiences..

B

This is all linked to the portal., but we also have experiences at the id,, and I want to show you that first. sure., let's take a look at those. so,. One of them is searching, live telemetry.

B

If you click on, that, will actually show you all of the telemetry we collect. so,. Remember I mentioned application. Insights is an app model level. Monitoring? yeah.

A

So,, what that means is, we actually go, and we.

B

Hook into all the requests coming in, and we see them individually., we look at every single dependency code,, you may call it like http calls,, sql calls,, and we collect the parameter series. As well..

B

We look at your exceptions,. We collect the call stacks.

B

A

With the profiling and debugging in.

B

The context of the scope of your application. so,, it's all connected to your code. and here,. You can see the various telemetry items that we have been collecting. for instance,, page views,, custom, events, there'll be traces of exception, dependencies..

B

Those are different types of telemetry that ai collects and then,.

B

Uploads to analytics for us to make sense of.

B

all of this is live and you can actually consume this from visual studio without ever leaving visual studio, id. so,. What's interesting about this is,.

A

I got to have this in editor experience..

A

Do I have to do anything special to get this??

A

Do I have a particular extension? I need to install,, or is it once I have visual studio,? I kind of just have this extension, and I just need to turn on application. Insights?.

B

So,, if you're using visual studio, 15 or 17,, this is already built-in.

B

okay.. All you do is just say: I want ai..

B

There are additional nuggets. You can download to get additional diagnostics, like extension, debugging., sure. and we'll slightly touch on that later..

B

But most of the core experiences are all built-in.

B

got it. and so also,.

A

With this experience, we have here,. Is this a purely local debugging experience or like.

A

For instance,, are these events stored in the cloud somewhere? or are they just local events?.

B

This is coming from the cloud.. This is actual live experience and that's what makes this so great., because you can be testing inter-loop as much as you want. yeah., but then, you put it out in production and it's a very different experience..

B

This lets you go and debug live, what's happening in production. and then,. You can build workloads on this,. I can do a proof test, load, and then I ran a couple of loads, and I'll. Show you afterwards that the data spiked.

B

sure. so, you'll, see all of this here.. What I like about this really cool is,.

A

I usually like to work with multiple monitors.

A

B

You know,, I have.

A

Something that's running visual studio.. I may have my web browser in one or.

B

A database or other monitoring tool in the other one.

A

So essentially, I could take this tab and put it in another monitor, right? and I could just kind of watch my telemetry that's coming. In.

B

A

Right. and as I continue my development experience?.

B

In fact,, let me show you something really cool.. We have this feature called,, so I'm going to jump into the portal. okay. called live metrics, and this is very close to your perfmon experience.

B

It's just this is perfmon for azure, right?.

A

B

Or in fact, any monitoring workload from azure.

B

And here, I'm actually monitoring this theory application and let me actually go and generate some load.

B

And, as I generate load, you'll see that it will show up, here, up. it'll show up here.

A

B

And if I were to create an exception by clicking on something that doesn't work,.

B

It'll show up there right away as well.

B

So, now you see exception rate just jumped.

B

Right. so,, I see on the right side, here,.

A

That it says, exception method., and it gives you the message and disconnect.

B

And you can actually go into this. name space going to see. and you can see the entire call. Stack. great. awesome.

A

So,, this is live monitoring of.

B

The production workload running in azure.

B

and so, is it essentially capturing all of my events?.

A

Or is there like some filter that I could apply to this to say, 'hey, only capture these particular types of events and tracers and such?' so,? That's actually a great question.

B

This is where this configuration file comes in,, and you don't need to figure this out. The first time you use it.. We actually set the right smart defaults for you. sure., but we have something called adaptive sampling..

B

It would be too expensive and too much overhead to collect a hundred percent of the requests coming in every single dependency. yeah.

A

But for a given request,.

B

You want to know everything that happened, usually. otherwise,. You cannot debug. sure. right. so,. What we do is we sample, and by default, we do five percent sampling, but we're adaptive in that.. If we see a burst of workloads,, we will actually throttle ourselves so that we don't impact your production. Throughputs.

B

got you. so,, it's almost like intelligence, sampling.

A

B

So,, it's not going to affect the performance of.

A

My application,, it's not going to get slower. yes., but still I'm able to get this rich, useful information.. That is exactly it.. Correct. got you. so, other things.

B

We're having here are code line integration,, which means I'm right in the code., and I've shown this to many developers,. They all get excited about this., it's actually much more than the portal, because this is where they sit most of the time and spend their time.

B

yeah. so, you can actually, on top of the methods,, can click on "code lines", and this is live integration with that data. Again. sure., if you see exceptions in there,, you'll, see this here as well,.

A

You'll be able to jump to the data specific.

B

To just that particular method, if you wanted to.

B

got it., there are many more things, individual food integration.. I didn't have time today to show you this.

B

for instance,. You can go and look at trends and see what the anomalies are.

B

for instance,. This could be an anomaly based on load.

B

So,, it will tell you that kind of exceptions in time, but could still be different from the expected count of neural seven., because with this exception,, usually you've got close before. sure.

A

And then you can go double click on.

B

That and try to diagnose this..

B

I actually wanted to show you the portal experiences, because they are far richer than what you can get in the visual studio and many developers don't realize that.

B

sure,. But I think this is really cool.

A

So essentially, we're getting some form of almost like anomaly: detection, right?.

B

Yes, there is anomaly detection in application: insights., it's detecting regressions against baseline,; it's also detecting outliers against the baseline nearby.

B

Got it. so based on that,.

A

Like once application insights detects this,, can I set up like alerts or notification?.

B

A

To let me know, 'hey,, the world is blowing up.' yes,, you can., you know we should react to this.

B

And one of the reasons I wanted to go to portal is we actually have smart detection, which is a loading mechanism that is on by default,? You don't have to do anything and it will alert you anytime. There are regressions against baselines on things like request, failure rates or durations.

B

all right. well,, let's head over to report and take a look. so,. There are different ways of doing this., one of them,. If you are in visual studio,, just click on this, all right?

B

or you just go and login on the portal and search for application, insights resource, either one will work., and what that will do is that it will load up the azure portal and will show you the application, insights, resource.

B

B

One of the really cool things in the portal and the reason I think the portal is being underutilized, is the ability to dashboard. all of the experiences you see. Here, you can pin on dashboards.

B

and you can actually show them, put this up on a big screen,. Have a team review this on a daily stand. Up, see all the metrics.

B

there's a lot of data in application, insights and in 20 minutes,. We can't possibly cover all of these at a high level.. We have a loading, so you can detect web,.

B

Give you ability to see life, what's happening on the surface, and I showed you a bit of that.

A

So,, let's talk about it, really quick. so,. This says two servers. yeah., so essentially,. I can know how many clients are connected to my web application at any given point in time?, clients will be users.

B

The servers are the actual backend instances.

B

okay., and in this case, there are only two instances. we could have 11,. We could have 150. yeah.. We are choosing application size to monitor our own internal workloads, including application insights., and if you want,, I can show you how this actually looks on a real big production, server. yeah, that will be great..

B

So in this case, I would go to a dashboard.

B

It is actually a typical dashboard. I use to start most of my demos as I jump into the various parts of the experience., so essentially,. What you did is you.

A

Went into application insights and you selected certain pieces of data that you were like,.

A

'well,, these are the things that are important to me,. I would have come this to my dashboard.' correct. so,. Every time you log into the portal,, you could select this dashboard. exactly.

B

And this is the view that you're going to see.

A

B

And this could be focused to from.

A

One app to multiple apps.: it doesn't really matter. absolutely.. Whatever your collection of information that you think is important,, you can put here.. That is absolutely true and you can actually.

B

Build dashboards out of multiple different data, sources,.

A

So they could point to both infrastructure,.

B

Networking and application layer at the same time,, security. right?, you could have full- are back this proper azure provider, resources.

A

So,, it's a really great way to collaborate as.

B

A team, or even just to go back to the same state., sure. so,. This is one of our instances that is monitoring the actual application. Insights, portal.

B

and you'll see it, a lot more data in here than is in our demo app. okay., and you can see there's about 100,000 users that have been using our service in the past 24 hours, with hunt and availability..

B

We actually rock both of them up.

B

right. so,. This is a live stream and this would be monitoring exactly what is happening in production, right, now. and if there were lsis,, people will be jumping on these, and using this area to go and debug application. Sites, in this case, front-end.

A

Got it. okay. so, different scales as much.

B

As you want,, you can throw with it. and the reason it scales,. We haven't talked about the magic in the books,, but it's all powered behind the scenes by something called kusto, which is application, analytics., right. so,. Let me show you a bit of that as well.

B

Sure., you can get to it from here, right into overview., and that will give you this opening experience or you can get it from any of these icons and they're sprinkled throughout the entire product.

B

so,. Wherever you go,, you will actually find these. and while I'm at it,, I will show you our new apm experience,.

B

This is something that lets you debug and analyze performance over time., okay. sort of in my previous life,, I spent 10 years at windows, working performance,, right?

B

yeah.. This was the golden dream for us being able to,.

B

At your fingertips, interactively ask for 92 percentile.

B

So that's, right now,, it's separate,, right? yeah.

A

And now, it's 90 percent up.

B

Be able to zoom in on any given section and the spikes that you see there..

B

That's actually me running a perf test, load.

B

so,, I ran,, I think about 500 users.

B

in fact,. Let me show you how well this is integrated.

B

Just go to "performance testing", and this will show you the recent perf test,.

B

And you can also schedule new ones. so,. I ran a couple today. so,. You run the perf test.

A

Inside of application insights,, this is not like a esp on that perf test or some external tool you created here and ran it here.

B

I use the user experience,, so application insights blade to do this., but remember, azure is actually a collection of lots of different services that are very uniform, ui and blog windows., all right. so, to the user, yes,. It looks like an application insights..

B

It's a separate infrastructure that is doing, in this case, visual studio lot of testing. got it., okay. and yet,. If I wanted to,, I can actually go and create a new one., so actually,, let's to that., to make sure that this is very believable in life,, I'm going to run another one..

B

If you'll ask me, what's your real tipping,, so I'll just go in, pin one.

B

So now, for a minute, he's going to do 250 users load.

A

Okay. so,, you set the duration, the test name and you schedule,, you send it to go, and it's pretty much about it.

B

So now that spike will show up and we can actually probably see this live..

B

This would be a good way to look at life.

A

And at any moment, now,, we should see like some spikes going up from that low test, that you--.

B

A

Scheduled to run. correct., that's interesting. yeah., and so, while we're taking a look at this.

A

so,, you mentioned that you used to work on windows, performance. so,. What kind of applications is application? Insights for?? Is it just for web apps,.

A

Or can I use it for desktop apps, for mobile apps??

A

Can I use it for iot devices, for instance, like what can I, you know,? Where can I use this? so technically,? You could use it for anything.

B

But our focus is on workloads running on servers.

A

Okay. and then the browser's connecting to those workloads.

A

okay. so,, I could profile my javascripts., you could. as we're waiting for this to load,.

B

Let me go and show you, there. Is.

B

When you go to the "performance" blade,.

B

By default,, that's not going to show you client information. in fact,, you see client topic,, customer browser. sure., but we have a separate blade here for browser.

B

and you can see what kind of page views were happening, where they're slow, was ajax call slow..

B

All of that is being jscript-injected with.

B

Jscript, that is being served from the backend to your client, to the browser., and you can double click into those as well,. If you want to.

A

Right. so,, when you showed the browser tab,.

A

The first thing I thought was there's a lot of times where I'm building applications and I'm sending this telemetry or this diagnostic data.

A

yeah., but sometimes,, it's more so just trying to figure out where my users are coming from,.

A

What browsers are they using,? What operating systems are they using.? So it's not just a matter of having.

B

Monitoring and performance and load testing.

A

But it can also get some rich information about like,, who are my users?? Who are my clients?? How many times do they come to my application and the features that they're using?

A

so really gets some really rich information on application? Insights.? You talked about users' analytics.,.

B

We have an entire offering for that as well.. It's currently in preview. okay., and you can find this under the "usage section" in the application. Insights menu.

B

got it.. It does look you segment by things such as operating systems,, browsers,, countries. sure.. It lets you see how well your retention is or not.

B

It lets you build funnels, so you can actually see your conversion rates.

B

Okay., probably my most favorite feature here is actually user flows., so I'll show you "funnels" first.

B

You can define whatever custom events.

B

You want, as you convert.

B

so,. This is really useful for this apps, because as, especially,, the product manager, you go and you focus on how many of the customers who came into the first blade, went to the second blade, went to the third blade.

A

B

Checked out, right?.

A

This is probably my most favorite user experience.

B

So what this apps is going to dynamically on the fly based on the telemetry data,, show you the navigations, through your entire app. nice. in this case,.

B

This is "overview" page,, which is this and.

B

People are navigating in different pages., so you actually,, you go and you see these different pages right. Here.

A

I noticed on the bottom in the little tabs,, it says: "sessions".

A

so,, that will be the unique session.. That's right.

B

That started in the beginning.. That's correct. showed, as it flowed through the application.

A

Yes., you can do this by users,.

B

But as the default, use by sessions- and this is very customizable.- I don't have time for this,, but we could have another session just on that.. Just on this. show how you can build user flows.

A

So, one of the things I'm thinking about,, as you show us all, these features is that you must store a lot of data to be able to generate some of these things..

A

We do. so,. One of the things I'm wondering about is like,. How long is this data actually stored? For??

A

The full retention is currently at 30 days.

B

30 days. okay. then,. You can store it for longer,, just pay for it per gigabyte. per gigabyte.. We're also throwing the first five gigs for free..

B

First five gigs. so,. I can essentially.

A

Get started using application.

B

A

Don't pay a single cent for it.

B

If you don't have enough workload,, if you sort seal the inner loop,, you probably will never even reach the five gigs. right. nice. okay.. Every month you get five gigs. got it.. So now, you see the workload. there's actually live, lotus happening in production and our.

B

Vsts workload generator is hitting it with 250, similar users. yes.. This is something you would do just before. You deploy it, for instance. right? yes.

A

There's something else you will do when you are.

B

Deploying. something you will do when you are monitoring this big, no name event around the conference, et cetera, et cetera.

B

sure.. Can we take a look at the analytics expert?

B

yes. so,? I just keep getting to it. that probably the best entry point in analytics, so it actually is not scary..

B

If you don't know what the language is, is to go through one of these icons, because what we did is the entire you actually see is built using the same kusto queries that you can write. Yourself. sure.. We made it possible for you to get to them.

B

so,. If you go to this blade and we say "view analytics", either one of these views, I can go, and I can just ask to be generated in application. Analytics.

B

It will give me the full query that we use ourselves to generate it. okay. so, that particular chart that we.

A

Just looked at, the one that you click on,.

B

It will show me the--.

A

It will-- the kusto query. in kusto,, being the query language that we use, require these metrics..

A

That is correct.. The reason it's called kusto is because.

B

Of the jacques cousteau,, which is explorer of the sea.

A

This big data is that sea of data.

B

Right. nice. so,, that's where this name came from., that's very cool. so,, it's the same exact histogram in this case, and this is probably one of the more complicated examples of a kusto query..

B

This would be a simple one,. It would be a trend.

B

That's the best way I find to get.

B

Into the language and start changing things like changing the time, range, changing the operation, names. so,, it's not a complicated language..

B

It took me about 20 minutes to pick it up,, actually.

B

A

Online, we have documentation, the one that we could look at.

B

The recent mono documentation is secured, very well looking at this.. If you go to akms and then kusto,.

B

A kusto settings both work.

B

there is a really good documentation in there.

A

Okay. awesome. then,, I can come in here,, learn.

A

About the different queries,, the parameters that I could use., you can also use the default getting.

B

Started experience of the application because it has.

B

so, if you're coming from scratch or using,.

B

There's another protocol log analytics, which is basically kusto as a service itself. okay.. They have a really good, getting solid experience here as well and they give you sample queries..

B

But if you really playing with data from ai,, it's sort.

B

Of much easier to just jump into our own queries and searching it, up. got it.

A

So on the left-hand side,, it looks like I have a section of, I guess, topics, or like what do we call those?.

B

So,, this is actually not topics,. This is data that you can work. With. application insights has its own well-defined, schema.

B

For the log analytics in this case, right?, so we are inside log analytics experience. Here. got it., you see the same [inaudible]. I talked about. "traces", "custom, events",, "page, views",, "request",, "dependencies", "exception", et cetera.. These are the tables of, in this case, "traces" and that's the schema for the "traces".

A

B

This would be "requests" and this is a schema of "requests". and then you can run queries on top of these data. Sets.

B

You can join them if you want to, you can, in fact,. That's what we do we'll look at these requests and then we join them by correlated activity, to see all the failures, all the dependencies that came as part of that.. I could actually show you that.

B

for instance,. If I were to go and look,, let's see something interesting.

B

This is different,, I don't want debug production.. Let me go back to that.

B

If we are going to go and look at say,, one operation like git index or something., and I.

B

Can look at various samples for that operation and just jump into one of them..

B

This was a slow one,. It was six seconds., probably not a good idea to load your home page in eight seconds..

B

You can actually see why it was slow. As well, we will show you,. That was a sequel problem.

B

in fact,. If you double click into this,, you will find out. This is a sequel problem, because I have overloaded this instance,.

B

It wasn't using the right tier. so for 60 connections, concurrent in just a time, out. sure. the fix would be just to upgrade tiers,, but you see this all double telemetry. For this whole operation.

B

so,. This is the only items..

B

We can run these queries for you and it will show you the related items..

B

It's going to be the same, query. nice..

B

I think there is a lot of discussion that we can have around each of these experiences.

A

B

There is ability here to go and get to.

B

The performance traces, so you can actually go and get to views like this..

B

This is what you use when you debug local applications like on desktop..

B

This is a performance analyzer.

A

Everyone of these profiles,.

B

Actually,, I haven't show you the profiles: yet.

B

so,. If I click on this button,, these profiles we're collecting in production using.

B

The same technology, making sense to them as you click "unpacking", it'll, show right. Now., I'm going to show you call stack.

A

Really? nice. yes. so, at that moment in time, you caught this particular call stack and so, wow.

A

So,, I could drill all the way down into this.

B

In the portal, never leaving the portal. nice., that's what I said, you don't actually have to have visual studio code on your machine. yes.. It could be a dev of skype.

B

and even if you're, a developer,, I mean, I've been a developer.. Where I went was debugging production, quality, postcode.

B

so,. I don't need to go and install their entire solutions, in this case,. I can see all of this straight from the portal.. That's great. for my [inaudible] ,.

A

I essentially have deep integrations that I can see the call stack,. I can see very detailed information about how long these methods took to execute when they're executed.

B

You will see both your code, as well as the dependencies you're making. okay. so, in this case, azure queue will spend quite a bit of time. in fact,. It will also upload different calls. yes. you'll, see exactly what's on [inaudible] path.

B

in this case,. It was because we weren't making any progress. Here, you were waiting up for this, right? got it. got it. got it.. If you want to dig down deeper, download the file., you have these rich views, right?, then you can go as deep as you want..

B

This is literally going down to context, which is so you can say,. 'this thread was blocking everything else, so this look was finished.' so,. This is actually the first time that you can debug.

B

Production workloads with this no repro,.

B

No excuse mentality,, which is you have evidence., there's nothing else. You need., it's all there. got it.. We brought you the same thing for both performance and for debugging.

B

So, there are exceptions. one of the features we noticed recently was exception, debugging and that's a really, really powerful, feature. so,. What happens is every time there is an exception happening, we will clone actual process and we'll let your process run. So it doesn't block.

B

then, on the shorter process, we're going to do a dump of your memory., we're going to zip it up,.

B

Send it back to the cloud,, make sense of it and show you in the portal, call stack with their labels. so, everything that happened,. So you have,.

A

Instance in time of what was the problem at that point in time- and I know you can inspect it and do some intelligent things. that is correct. yes. so,, one of demos. I do in.

B

This application is that it's being validation on zip, codes., yes. so,. That was this code, and you find out that the reason.

B

This failed is because someone entered the canadian zip code instead of the us, zip code. sure.

A

I can see this because I have the local variables.

B

I have the entire state in the back extension.

A

Nice., that's awesome. so,. We definitely have to have you come back on another time, because I feel.

B

Like there's so much stuff to talk, about.

A

But this is a really good start in terms of just people, understanding what exactly they can do with application insights. yes. so,. I highly recommend that you try it.

B

Out because it's free to start with.

A

It gives you a lot of telemetry for free to play with and.

B

It brings to the more insights than anything I've ever seen on a local box., that's cool,, man., so again,.

A

Thank you so much,, michael for coming along. and here,. We have another episode of the on .net show..

A

We learned all about application. Insights.