Using Azure Monitor as part of your development workflow

Posted: (EET/GMT+2)

 

If your application is running in Azure, you already have access to logs and metrics. The useful step is feeding that data back into development.

Azure Monitor and Application Insights make it easy to see what is happening in production. The challenge is turning that information into something actionable for developers. In the DevOps spirit, the feedback loop is crucial.

To get started, don't attempt to get feedback from everything. Instead, start with a few high-value signals instead. For instance:

  • Failed requests and exceptions
  • Slow requests
  • Dependency failures (database, APIs)
  • Resource spikes (CPU, memory).

These usually point directly to real problems affecting users. Similarly, you don't need a highly structured process to select which items to pick. Often, a simple review workflow is enough for most applications. Something like this:

  1. Open Application Insights
  2. Go to "Failures"
  3. Check recurring exceptions
  4. Go to "Performance"
  5. Find slow requests
  6. Inspect one (or just a few) concrete example(s).

Once you have the data and an idea for how to improve, the important part is what happens next. When you find something, follow a simple process:

  • Create a work item for recurring errors or slow endpoints.
  • Include a link, query, or screenshot from Application Insights.
  • Fix the issue in code or configuration.
  • Verify after deployment that the metric improves.

This closes the loop between production and development. Without that step, monitoring data tends to be ignored until something breaks.

Let's take a simple example. Assume an API endpoint is consistently slow in production. You see it in the Performance view. You create a work item, optimize the database query, deploy, and then confirm that the request duration drops.

That is the whole idea. Production telemetry becomes part of your normal development workflow.

So the practical takeaway is simple: use Azure Monitor to see what is happening, and turn those findings into concrete development tasks.