debug-pipeline

To diagnose a failed Harness pipeline execution, you provide an execution_id or pipeline_id to generate a structured report identifying bottlenecks, errors, and logs. This diagnostic scan applies to failed, successful, or in-progress runs.

A Harness pipeline diagnostic report includes stage and step breakdowns, execution timing, bottlenecks, and detailed failure information. You can also request logs, YAML definitions, and raw payloads for a comprehensive deep dive analysis.

You get pipeline execution logs by requesting a deep dive diagnosis through the harness_diagnose MCP tool. This fetches execution data, logs, and pipeline definitions to provide a comprehensive analysis of your pipeline run.

Yes, you need Harness MCP v2 and the harness_diagnose tool to run the pipeline execution analysis. These dependencies are required to fetch the execution data, logs, and pipeline definitions needed for the diagnostic report.

Yes, you can diagnose a pipeline that is currently in progress. The diagnostic scan applies to any Harness pipeline run, including failed, successful, or in-progress executions, to identify bottlenecks and errors.

The best way to spot trends in project pipeline failures is to use the context-aware health overview. This feature fetches project health data to help you identify recurring failures and trends across your Harness pipelines.