The Kubernetes Troubleshooting Agent performs comprehensive analysis of active issues to help make troubleshooting faster.
Using Dynatrace Intelligence and out-of-the-box alerts, the Kubernetes Troubleshooting Agent:
To use the Kubernetes Troubleshooting Agent, you must deploy a workflow that is included in the product as a preconfigured workflow template. For details, see Agentic workflows (Preview).
To set up the agent
Go to
Workflows and select Workflow.
From the Create workflow from template list that opens, select Kubernetes Troubleshooting Agent.
Install all the required apps (as needed), define the connections, and select Create workflow.
Optional To configure the JSON workflow input, select and scroll down to the Workflow input and result and Default input sections.
Update the generateRecommendation parameter value to true to automatically enrich any Kubernetes problem with the output from the agent, or to false to stop the agent output. The default value is true.
Update the generatePullRequest parameter value to true to extract the YAML from the agent's suggestions and create a pull request with the changes, or to false to disable this functionality. The default value is false.
To enable this functionality, the create_pull_request_1 step must be properly configured in your Git repository.

Test the workflow with a manual run by selecting Run.
If the test works as expected, no additional action is needed. The workflow will integrate AI-driven diagnostics into your Kubernetes environment and displays results in
Problems.

Kubernetes