Konflux Its Analyzer

Konflux ITS Failure Analyzer

Overview

Analyze failed Konflux integration test scenario (ITS) PipelineRuns using KubeArchive. Retrieves archived PipelineRuns, identifies failed tasks, fetches logs, and produces root cause analysis reports.

Namespaces: If the user does not specify a namespace, use one of:

• ai-tenant — RHAIIS components
• rhel-ai-tenant — RHEL AI components

Prerequisites: KUBECONFIG and KUBEARCHIVE_HOST env vars must be set.

IMPORTANT:

• Do NOT run any prerequisite checks (no command -v, no env var checks, no install scripts). Run the scripts directly — they validate dependencies internally and print install instructions if something is missing.
• Do NOT append 2>&1 to script commands — the scripts handle their own stderr output.

Scripts

get_failed_pipelineruns.sh

Lists failed ITS PipelineRuns within a time period.

Usage:

${CLAUDE_SKILL_DIR}/scripts/get_failed_pipelineruns.sh   [--human]

Arguments:

| Position | Argument | Description | |----------|----------|-------------| | 1 | ` | Kubernetes namespace | | 2 | ` | Time period (see formats below) |

Time specification formats:

| Format | Example | Description | |--------|---------|-------------| | Single date | 2026-04-16 | All failures on that date | | Date range | 2026-04-14..2026-04-16 | Failures from start to end date | | Relative hours | 4h | Last 4 hours | | Relative days | 2d | Last 2 days | | Relative weeks | 1w | Last week |

Options:

| Option | Description | |--------|-------------| | --human | Table output instead of JSON |

Examples:

${CLAUDE_SKILL_DIR}/scripts/get_failed_pipelineruns.sh ai-tenant 2026-04-16
${CLAUDE_SKILL_DIR}/scripts/get_failed_pipelineruns.sh ai-tenant 2026-04-14..2026-04-16 --human
${CLAUDE_SKILL_DIR}/scripts/get_failed_pipelineruns.sh ai-tenant 4h

Output (JSON):

[
  {
    "name": "rhaiis-test-vllm-podman-cuda-x86-64-msrt7",
    "namespace": "ai-tenant",
    "created": "2026-04-16T23:49:09Z",
    "application": "rhaiis",
    "component": "rhaiis-cuda-ubi9",
    "scenario": "rhaiis-test-vllm-podman-cuda-x86-64",
    "optional": "true",
    "event_type": "Merge Request",
    "reason": "Failed"
  }
]

---

analyze_its_failure.sh

Deep-analyzes a single failed PipelineRun. Examines TaskRun statuses, fetches failed task logs, and produces a root cause analysis.

Usage:

${CLAUDE_SKILL_DIR}/scripts/analyze_its_failure.sh   [OPTIONS]

Arguments:

| Position | Argument | Description | |----------|----------|-------------| | 1 | ` | Kubernetes namespace | | 2 | ` | Name of the failed PipelineRun |

Options:

| Option | Description | |--------|-------------| | --human | Human-readable output instead of JSON |

Examples:

${CLAUDE_SKILL_DIR}/scripts/analyze_its_failure.sh ai-tenant rhaiis-test-vllm-podman-neuron-x86-64-pqr7h --human
${CLAUDE_SKILL_DIR}/scripts/analyze_its_failure.sh ai-tenant rhaiis-test-vllm-podman-cuda-x86-64-msrt7

Output (JSON):

{
  "pipelinerun": "rhaiis-test-vllm-podman-neuron-x86-64-pqr7h",
  "namespace": "ai-tenant",
  "status": "Failed",
  "summary": "Tasks Completed: 7 (Failed: 1, Cancelled 0), Skipped: 6",
  "event_type": "Merge Request",
  "commit_sha": "abc123...",
  "application": "rhaiis",
  "component": "rhaiis-neuron-ubi9",
  "scenario": "rhaiis-test-vllm-podman-neuron-x86-64",
  "optional": "true",
  "repo_url": "https://gitlab.com/redhat/rhel-ai/rhaiis/containers",
  "merge_request": "https://gitlab.com/redhat/rhel-ai/rhaiis/containers/-/merge_requests/351",
  "log_url": "https://konflux-ui.apps.../pipelinerun/...",
  "failed_tasks": [
    {
      "task": "test-inference",
      "step": "unnamed-0",
      "exit_code": 1
    }
  ],
  "analysis": "Human-readable root cause summary"
}

Common Workflows

Workflow 1: Find and Analyze Failures for a Date

# Step 1: List failed PipelineRuns
${CLAUDE_SKILL_DIR}/scripts/get_failed_pipelineruns.sh ai-tenant 2026-04-16 --human

# Step 2: Analyze a specific failure
${CLAUDE_SKILL_DIR}/scripts/analyze_its_failure.sh ai-tenant rhaiis-test-vllm-podman-cuda-x86-64-msrt7 --human

Workflow 2: Recent Failures

# Check last 4 hours
${CLAUDE_SKILL_DIR}/scripts/get_failed_pipelineruns.sh ai-tenant 4h --human

# Check last week
${CLAUDE_SKILL_DIR}/scripts/get_failed_pipelineruns.sh ai-tenant 1w

Error Handling

| Scenario | Behavior | |----------|----------| | PipelineRun not found | Exit 1 with error JSON | | PipelineRun not failed | Exit 0 with status info | | KUBECONFIG not set | Exit 1 with error | | KUBEARCHIVE_HOST not set | Exit 1 with error | | kubectl-ka not installed | Exit 1 with install hint | | Logs unauthorized | Analysis notes "Logs unavailable" | | No failed tasks found | Analysis notes "No failed tasks" |

Dependencies

Required: kubectl-ka, jq — installed via tools/*/install.sh

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: aipcc-cicd
• Source: aipcc-cicd/claudio-skills
• License: Apache-2.0

Install and usage instructions live in the source repository linked above.