Epic Goal

Deliver a productized Tekton Assist service that leverages LLM-based analysis to explain failed Pipelines/TaskRuns and guide remediation. The service will be deployable via Helm, integrated with the Tekton CLI (tkn), and supported under the openshift-pipelines org.

Why is this important?

• today, diagnosing failed Pipelines/TaskRuns requires deep Tekton expertise and manual inspection of logs.

• Tekton Assist reduces time-to-triage by providing actionable root cause analysis and remediation guidance directly to users.

• Improves developer experience and operator efficiency by integrating into existing Tekton tools (tkn, Helm).

• Aligns with OpenShift Pipelines’ goal of delivering AI-powered developer productivity features.

Scenarios

• A developer runs a Pipeline that fails due to missing secrets.

• • tkn taskrun explain my-run outputs: 

    ❌ Step 'build' failed due to missing secret 'docker-creds' Suggested fix: Create secret 'docker-creds' in namespace 'ci-tools'
    

     
     

• An operator installs Tekton Assist via Helm with an OpenAI key and custom model settings.

• The service deploys successfully and is reachable at a cluster endpoint.

• A TaskRun fails due to an invalid image reference.

• Tekton Assist provides root cause: "Image 'quay.io/foo/bar:latest' not found"

• •  
• Suggests checking registry permissions and image existence.
•  

Acceptance Criteria (Mandatory)

• ✅ CI MUST be running successfully with tests automated for Tekton Assist service and CLI integration.

• ✅ Helm chart available for Tekton Assist with documented configuration options.

• ✅ tkn pipelinerun explain and tkn taskrun explain produce structured, human-readable analysis.

• ✅ Release Technical Enablement: Provide necessary docs (install guide, CLI usage, troubleshooting).

• ✅ Tekton Assist service productized in openshift-pipelines org with Konflux CI integration.

Dependencies (internal and external)

1. ...

Previous Work (Optional):

1. …

Open questions::

1. …

Done Checklist

• Acceptance criteria are met
• Non-functional properties of the Feature have been validated (such as performance, resource, UX, security or privacy aspects)
• User Journey automation is delivered
• Support and SRE teams are provided with enough skills to support the feature in production environment