HomeBlogTexas Ranger

Previous arcticle: Application Insights End-To-End

The various APM-use-cases we described in Application Performance Management Overview are meant for different target groups: Developers, architects, devOps, ops as well as business departments. They usually use different tools to accomplish their tasks. Developers might use profilers whereas business oriented people need something like Google Anlaytics. Different tools may cause silos instead of an integrated communication flow. Application Insights can't of course satisfy all communication needs across the the application lifecycle but it can help a lot. One immediate improvement in cross-team-communication are charts and dashboards. They provide a quick overview and allow for drilling down to the details.

In our previous post we've shown how to add end-to-end-metrics to our APM, e.g. for a search performed by an end-user. As a first step towards our dashboard we will create a chart that visualizes the custom event metric duration.

"Metrics Exporer"



This chart is not only useful for developers and architects but also for business-people who are interested in the performance of the user interaction/transaction "search".

Next, we put together a Dashboard consisting of following parts:

  • The overview timeline: Visualize high level info: Server response times, page view load times, the number of server requests and failed requests.
  • Server-KPIs: CPU consumption, I/O rate, exception rate and HTTP request rate.
  • The chart we just created.

"Dashboard"


This might be a cross-team dashboard: high-level, yet providing drill-downs to the details. To create a new Dashboard go to the overview page of the azure portal and click + New Dashboard in the menu bar. To add an element to a dashboard, select/activate the dashboard, go to the respective element (e.g. chart) and select the "pin" icon in its menu.

In our following posts we are going to show how to use AI to discover anti-patterns early and how to put anti-pattern-indicators on our dashboard.