We are running 200 - 300 UI test cases from an Azure pipeline. All the test cases are running on self hosted Windows agents, where all the dependencies are installed. Chrome browser, vstest, etc.
Issue is when we use parallelism multi agent from Azure pipeline, the pipeline runs for a long time without throwing an issue. Sometimes browser is getting stuck while cases are running. Sometime no response status appears in the pipeline logs. We are using 32bit, 8cpus machine.
Please suggest the best method to overcome this problem.
You need to analyze why the UI test is stuck. You can share the logs that you think are problematic in your question. Personal experience for this issue is generally because the mode of the private agent is not set correctly.
When we create the private agent, we could set two modes
Interactivevs.service. If we set the agent as service for the UI test, the build/test will be blocked. Since the tested UI interface cannot pop up in this mode. We need to generate the private agent as Interactive mode:Interactive vs. service