DevOps > CI + CD. A web developer's introduction to Application Insights

Editor's Notes

• #7: Objective: To show that the DevOps cycle is incomplete with just CI&CD and that Monitoring/Telemetry is an integral part of the cycle. Notes: In a typical DevOps cycle, you come up with business objectives, goals, and a backlog for your solution. From there, you enter a development/test phase, which is facilitated by the use of Continuous Integration (CI) tools & techniques. From there, you release the product for use, relying on Continuous Delivery (CD) tools. But then what? How do you objectively connect your product releases to making choices about goals & objectives in the business? That’s where the Monitoring & Learning phase comes in. In this case, telemetry-gathering tools can provide objective information about your application usage and both possible successes and potential issues. Within a telemetry pipeline, there are typically 3 phases – Ingest where the data is gathered; Explore, where the data is analyzed and interpreted; and Report, where the data is shared with project participants and stakeholders.
• #9: Objective: To show the high-level features that make up Application Insights Notes: The Application Insights service is made up of several different pieces which can be brought together to address the 3 key phases of a Telemetry Pipeline.
• #10: Objective: Provide a quick overview/tour of Application Insights Notes: Show Application Insights applied at runtime over a running application. See telemetry; focus on Search, Metrics blades.
• #11: Objective: Show the different capabilities that you can leverage depending on whether the app is instrumented at runtime, buildtime, or both. Notes: Applying App Insights at runtime is quick and easy, but can be limited in terms of what telemetry is displayed. Likewise, if you only leverage the build-time approach, information is only collected in-proc. Leveraging both solutions together provides the most robust feature combination. Note that for Run-Time support outside of app services, the IIS Status Monitor can also be installed. Link: Instrumenting Web Apps at runtime: https://docs.microsoft.com/en-us/azure/application-insights/app-insights-monitor-performance-live-website-now
• #12: Objective: Illustrate how Visual Studio tooling can make it simple to configure Application Insights at built-time. Notes: With Visual Studio, adding Application Insights to your application is straightforward. First, select the “Configure Application Insights” from the project file’s context menu. Then specify the Application Insights resource in Azure to use. Press the “Register” button to create the Azure resource if necessary and then to add the necessary configuration to the Visual Studio project.
• #13: Objective: Discuss the different kinds of custom telemetry that can be reported with the App Insights SDK Notes: In addition to the telemetry that is captured by default with the App Insights Modules, you can also use the App Insights SDK to log telemetry entries. There are several different kinds of telemetry you can track: PageView is used to track custom page views (sometimes things like tab pages or other SPA activities should be considered “Page Views”) Requests are similar to Page Views, but are more aimed at Web API requests Events mark the occurrence of something notable, often used to count feature usage or user activity Metrics light up in the Metrics Explorer, and can be used to track aggregated events (count and sum, min and max, etc.) Exceptions track exception details and are especially useful in Global Exception Handlers if the app pre-processes exception details before returning them to end-users. Traces track general messages. These items are emitted by System.Diagnostics Trace calls when the Application Insights TraceListener is used Dependencies are used to track response times and success rates for calls to external items. They can be used if you do not want to install the collection agent, or to catch calls the agent does not catch.
• #14: Objective: To show that data can be collected into Application Insights from a variety of application types Notes: Application Insights is not just a technology for ASP.NET applications. There are both officially supported and community created SDKs and Frameworks for all kinds of apps and runtimes. If you’re using ASP.NET Core, the support is limited, but there’s ongoing work on the SDK to move closer to parity with “classic” ASP.NET (see the issues in the repo: https://github.com/Microsoft/ApplicationInsights-aspnetcore) Java server apps are also supported, both with runtime instrumentation and with a build-time SDK Node.js servers are also supported with an npm package; there are also client-side SDKs and frameworks that can be used for SPA apps. You can also set a variable in .NET app settings and it will automatically include a JavaScript snippet to do page performance reporting Windows Services & Desktop apps are supported with the .NET SDKs; though it is recommended that desktop and mobile client apps use HockeyApp and its successor Visual Studio Mobile Center Finally, Android, iOS, UWP, WPF and OSX apps can use HockeyApp as their monitoring tool, and you can set up an App Insights resource that is configured as a bridge between the two systems, allowing you to report data to App Insights from these apps.
• #15: Objective: Show how the Visual Studio tooling configures the app to use Application Insights. Also show how the various custom trace messages can contribute to telemetry collection Notes: Go through the configuration process. It can be much faster to connect to a pre-existing App Insights resource. Once complete, show the various changes, including: Nuget packages, ApplicationInisghts.config file, web.config changes (including Module AND TraceListener) Publish and run the app and generate new data. Show additional parts of the telemetry lit up (Application Map with dependencies, Live Metrics, detailed exception information.) Show the code for generating events.
• #16: Objective: Discuss ways that Application Insights data can be explored in the Azure Portal Notes: The Azure Portal offers a long and growing list of different ways to explore Application Insights data. Some highlights include: An interactive Application Map that shows the relationship between different parts of your application as they communicate between each other. A Live Metrics Stream viewer that allows you to view real-time telemetry flowing from your applications Metrics Explorer where you can chart select application metrics and aggregations, select time ranges, and drill into the data that generated the displayed measurements. Interactive Telemetry Search which you can use to explore the gathered telemetry, filter and search, and explore across items related through common operations and sessions. Set up application availability probes and visualize, analyze, and drill into the results of these probes Several User displays where you can explore user activity, site retention, and user session information. Several other different ways of exploring your application data.
• #17: Objective: To discuss the ways in which Application Insights integrates with the Visual Studio IDE Notes: Visual Studio also includes integration with Application Insights so you can get rich insights into your application’s behavior and performance without leaving the Visual Studio IDE. This also includes analyzing telemetry during development where the data has not yet been sent to the Azure Portal. Places where Application Insights is integrated with Visual Studio include: A Telemetry Search window which provides functionality similar to that which is available in the Telemetry Search blade in the Azure Portal, but also includes clicking through to the underlying code. A Telemetry Trends window that provides a graphical display of how an application’s telemetry events change over time, which can help pinpoint large or emerging anomalies. Integration of the Application Insights telemetry reporting into the Visual Studio console window, so you can examine telemetry information similar to trace messages while an application is attached to the Visual Studio debugger.
• #18: Objective: Present the new Snapshot Debugging feature Notes: TODO
• #19: Objective: Present the Application Insights Analytics Portal Notes: The Application Insights Analytics Portal provides more direct ad-hoc interaction with the application telemetry tables. It includes a query editor with Intellisense, and a data display region query results can be displayed in either tabular or graphical form, and machine learning operators and user interface elements to help identify anomalies or trends in the data.
• #20: Objective: Show how the various exploration tools can be used with the Application Insights telemetry Notes: Show the telemetry starting in Visual Studio, go to the Portal, then over to Analytics
• #21: Objective: To discuss how Application Insights telemetry can be reported to users by displaying Analytics query results as components of Azure Portal dashboards. Notes: One way you can share the insights you gain while exploring your Application Insights Telemetry data, is by exporting Application Insights Analytics query results as elements within an Azure Portal dashboard. This way, you can create complex charts that highlight key application information in the same place as you display the other system health information that the Azure Portal makes available.
• #22: Objective: To discuss several different ways that Application Insights telemetry can be exported in order to be displayed and consumed by other tools. Notes: There are several options available for exporting data from Application Insights so it can be consumed by other applications.: There are several different ways that data can be exported for use in PowerBI. Power BI now includes a Power BI Adapter for Application Insights that can be used to setup a complete dashboard of telemetry within Power BI. You can also export Analytics Queries as M text (M is the Power BI formula language) and import it into Power BI as a dataset, which can be used to create individual visualizations within Power BI. Finally, you can consume data published with the Continuous Export feature which will be discussed shortly as a dynamic input to Power BI. More information: https://docs.microsoft.com/en-us/azure/application-insights/app-insights-export-power-bi Telemetry data can also be exported as Excel data, either as the complete table or as the tabular result of a given query. Export to Azure Operations Management Suite via the Application Insights Connector in OMS Log Analytics Finally, Application Insights includes a Continuous Export feature which allows you to create a job that will continuously output a copy of your Power BI data into an Azure Storage account of your choosing. This data can be consumed by several Azure or on-premises services to perform further analysis and processing of the application telemetry
• #23: Objective: To discuss the Alert options provided by Application Insights as another reporting mechanism Notes: In addition to exporting the Application Insights data, it is also possible to generate alerts based on your telemetry and trends that may emerge from the data. Metric Alerts can be created based on Metric Telemetry conditions, such as a threshold of a number of occurrences or a value range. These alerts can either send emails to system administrators or custom addresses, as well as trigger a configured webhook, or trigger the execution of an Azure App Service Logic App. Availability tests that are configured in the Azure Portal can also be tailored to generate emails or trigger webhooks based on how responsive an application is to these probes. Application Insights also includes several “Smart Alerts” – alerts created by default that look for certain trends in your application’s behavior and optionally provide email notifications when those trends emerge.
• #24: Objective: Show how the various reporting tools over the Application Insights data can be used to show application health Notes:
• #25: Objective: To review the pricing and performance characteristics of Application Insights Notes: The Application Insights Data Retention Policy is 90 days. Basic – pay based on amount of telemetry data with a 1GB base allowance, and then $2.30 per GB after that. Enterprise – pay based on the number of servers/PaaS instances that run the App generating telemetry A “Node” is a server generating telemetry (what about JS clients?) Continuous export – There are “free” manual exports available (export for Excel in UI, export in Analytics, ingest in Power BI, use REST API) Push to a Blob Storage Account Choose the data types to be exported Includes raw telemetry, plus location (calculated from incoming IP address)
• #26: Objective: To suggest some additional sources for obtaining Application Insights information Notes: