Analyze load test results using the test run results dashboard

In this article, you learn how to use the comprehensive results dashboard provided by Azure Load Testing to analyze the results of your load test runs. This dashboard presents key performance statistics, AI generated actionable insights, client-side metrics, server-side metrics, etc. to help you evaluate the performance and reliability of your application under load. You can quickly understand whether your application meets the performance expectations. You can also do deeper analyses and troubleshooting by slicing and dicing the data using the rich interactive features.

The test run results dashboard is available for load tests run from any of the interfaces like Azure portal, Az CLI, REST APIs, Azure SDKs, Visual Studio Code extension, or CI/CD pipelines. This article introduces the main sections of the dashboard, describes information you can find in each section, and explains how you can use the information for load test result analysis.

The results dashboard contains the following sections:

Prerequisites

• An Azure account with an active subscription. If you don't have an Azure subscription, create a free account before you begin.
• An Azure load testing resource that has a completed test run. If you need to create an Azure load testing resource, see Create and run a load test.

Test run details

In this section, you can see the status of the test run along with important information about the test run. For test runs in progress, it indicates the current state of the test run.

A test run that is currently in progress can be in one of the following states:

After the test run execution is completed, you can view the final status of the test run along with two sub states. The first sub state indicates whether the test run execution completed as expected. The second sub state indicates whether the performance criteria defined for the test are met.

After the completion of test run execution, a test run will be in one of these terminal states:

In addition to the run status, you can see important information about the test run like Start time, End time, Duration, Engine instances, Virtual users (Max), Virtual user hours, Test run ID, and Baseline.

AI insights

In this section, you can see AI-powered actionable insights to get a quick summary of what happened during the test run and the performance of your application. You can also see detailed insights on potential bottlenecks and some recommendations to fix the identified bottlenecks.

See Analyze test results using AI to learn more about AI-powered actionable insights.

Summary statistics

This section provides you an at-a-glance view of the performance of your application during the load test. You can see

• The total number of requests sent
• Duration of the load test
• 90th percentile response time
• Error percentage
• Throughput.

If you configured a test run as a baseline, you can also see a comparison of these metrics with the metrics from the baseline run.

If your test scenario contains multiple requests or samplers, you can view the summary statistics at a sampler level under Sampler statistics. Optionally, you can choose a suitable response time aggregation.

For multi-region load tests, you can view the comparison of performance metrics across regions and easily find out the regions that are performing better and the ones that aren't.

Test criteria

If you defined failure criteria for the test run, you can view the results of the test criteria evaluation in this section.

For each of the test criteria defined, you see a card that shows:

• The performance metric like response time, error percentage etc.,
• The threshold value defined
• The actual value observed during the test run
• The result of the criteria evaluation.

If the failure criteria are defined on an app component metric, a link to view the Azure Monitor graph of the specific metric is also available.

Client-side metrics

This section provides a graphical view of how the client-side performance metrics like response time, throughput, and errors vary with the load over the duration of the load test. These graphs are populated live when the test run is in progress. Live graphs enable you to analyze the load test results when the test run is in progress. You can analyze the graphs after the test run completion as well.

The client-side metrics are available at an individual sampler or request level, and at an aggregate level. You can use the filters to slice and dice the metrics as required and derive conclusions on the performance observed during the load test. For example, you can view only the 500 errors of one specific request by using the Requests and Error Type filters. Similarly, you can view the maximum response in one specific region by using the Region and Aggregation filters.

Server-side metrics

If you configured app components for monitoring, this section provides a graphical view of the resource metrics from the selected app components like Azure App Service, Azure Cosmos DB, etc. You can correlate these metrics with the client-side metrics to identify potential bottleneck components.

By default, Azure Load Testing shows the resource metrics that are most relevant to evaluate app performance. For example, for an App Service plan, CPU percentage and Memory percentage metrics are shown by default. Optionally, you can select the metrics and aggregations of your choice by selecting 'Configure server side metrics'.

Engine health metrics

In this section, you see a graphical view of resource metrics of the load generating engines:

• CPU percentage
• Memory percentage
• Network bytes per second
• Number of virtual users.

You can monitor resource metrics of the test engine instances to make sure that the test engine instances, themselves aren't a performance bottleneck.

For more information on engine health metrics, see monitor engine instance metrics

Next steps

• Analyze test results using AI
• Diagnose failing load tests
• Identify performance bottlenecks