While the App Center is great there is one glaring omission at the moment. It doesn’t cater for traditional web apps. But before there was the App Center there was, and still is, Application Insights Telemetry. Adding telemetry is very easy. From within Visual Studio select Add Application Insights Telemetry from the project menu and follow the wizard. When installed you’ll find a connection to Application Insights services, the insights package added, and an instrumentationKey added to appsettings.json. If your web app happens to be an MVC app you’ll find a reference to Application Insights Javascript snippet added to Layout.cshtml. The caveat with this very simple set up is that your Visual Studio subscription credentials are the same as, or linked with your Azure subscription. If not you will be prompted to supply them. If you are using a JavaScript framework like Angular or VueJs there are Application Insights NPM packages you can install. So, what does the telemetry tell us?

If we go to Azure and then select application insights and then pick our telemetry app which has been automatically created based on our Visual Studio project, we will get an overview dashboard. This shows us failed requests, server requests, server response time and availability. 

Now if we go to our app and do some navigating and then select live metrics stream. Let's just wait for that to boot up - here we go. If we click on a few links we can see the live data coming through. This is all well and good but what about if we have a failure. So, I’ll initiate a failed page request by asking for a non-existent path. We can see the failures coming through both graphically, and on the far right we can see the requested path that is causing the failure.

To get notified of failures we must create an alert. We do this by go down to the alerts menu item and then click create new alert. We must start with a condition which is going to trigger the alert. As you can see there's plenty of options here to choose from, but I'm interested in failed requests so I'm going to select that. We can see the number of failed requests so far. I’ll set the threshold to one and change the granularity to one minute for the purposes of demonstration. The next thing I need to do is create an Action Group. I'll name the group email failed requests, with short name email 404. You have a lot of action types to choose from, which is great from an integration point of view, but I'm just going to stick with an email message just to begin with. I’ll just enter the email details. OK and we'll just give that group a name. I need to select the group that I've just created this is a very important step, and if you don't the alerts won't work. And give the alert rule a name. I should point out that just below the description field there is a drop-down list where you can set the severity level. As you can see there have been 4 alerts in total.

Here we have an example of the email alert that gets sent by the telemetry service. On the right is the action type drop down list showing the other integrations you can enable. Let’s look at how you can turn Application Insight alerts into issues within Azure DevOps.