Complete Findings with Screenshots and Technical Details

Error Reproduction

Successfully reproduced the 500/404 error using automated Playwright tests with the following approach:

1. Initial workflow abort: ✅ 200 OK - Successfully aborted
2. "Run again" workflow abort: ❌ 500 Internal Server Error with 404 Not Found

Test Execution Details

• Test File: ui_tests/parallel/orchestrator_tests/userOnBoardingWorkflowAbortComprehensive.spec.ts
• Environment: OCP Edge environment (https://backstage-backstage-rhdh-operator.apps.ocp-edge73-0.qe.lab.redhat.com/)
• Workflow Type: User Onboarding workflow
• Test Strategy: Loop through multiple "run again and abort" attempts with incrementally increasing wait times

API Interception Results

The test intercepted and validated all abort API calls:

Attempt 1 (Initial abort): ✅ 200 OK

• Response: "Workflow instance {id} successfully aborted"

Attempt 2 (After "Run again"): ❌ 500 Internal Server Error

• Root cause: 404 Not Found from internal service
• Internal service call: {{http://user-onboarding.rhdh-operator/management/processes/user-onboarding/instances/ {id}
  }}

Key Technical Insights

1. Timing Factor: The error occurs consistently on the second attempt (after 2-second wait) but not on subsequent attempts with longer waits
2. State Management Issue: The internal service appears to lose track of workflow instances when using "Run again"
3. Intermittent Nature: The error is reproducible but not 100% consistent, suggesting a race condition or timing dependency

API Response Data from Test Execution

Successful Abort Response (200 OK):

Request URL: https://backstage-backstage-rhdh-operator.apps.ocp-edge73-0.qe.lab.redhat.com/api/orchestrator/v2/workflows/instances/{id}/abort
Request Method: DELETE
Status Code: 200 OK
Response Body: "Workflow instance {id} successfully aborted"

Failed Abort Response (500 Internal Server Error):

{
    "error": {
        "name": "Error",
        "message": "HTTP DELETE request to http://user-onboarding.rhdh-operator/management/processes/user-onboarding/instances/df1eacdc-cba3-48f8-adc9-8ac20e65eeb7 failed.\nStatus Code: 404\nStatus Text: Not Found"
    },
    "request": {
        "method": "DELETE",
        "url": "/v2/workflows/instances/df1eacdc-cba3-48f8-adc9-8ac20e65eeb7/abort"
    },
    "response": {
        "statusCode": 500
    }
}

Network Request Details

Abort API Call Pattern:

• *Endpoint: http://user-onboarding.rhdh-operator/management/processes/user-onboarding/instances/df1eacdc-cba3-48f8-adc9-8ac20e65eeb7*  
• Method: DELETE
• Headers: Standard authentication and content-type headers
• Response Time: Varies based on timing (2s wait = 404 error, shorter or longer waits = 200 OK)

Error Reproduction Pattern

Consistent Failure Point:

• Attempt 1: ✅ 200 OK (immediate abort)
• Attempt 2: ❌ 500 Internal Server Error (after 2s wait + "Run again")

This pattern confirms the issue is specifically related to the "Run again" functionality and timing of the abort attempt.

Test Code Highlights

The test successfully:

• Navigates to User Onboarding workflow
• Fills out workflow form fields
• Starts workflow execution
• Aborts running workflow
• Validates abort success
• Uses "Run again" button
• Repeats the entire flow multiple times
• Intercepts and validates all API responses

Screenshots and Evidence

The test captured multiple screenshots during execution:

• Workflow form completion
• Workflow running state
• Abort confirmation dialogs
• Aborted workflow status
• "Run again" button visibility
• Error states and confirmations

Root Cause Analysis

The 404 error suggests that when using "Run again", the internal service user-onboarding.rhdh-operator cannot locate the workflow instance, possibly due to:

• Instance ID mismatch between orchestrator and internal service
• State synchronization issues between services
• Timing-related race conditions in instance registration
• Database or cache inconsistencies

Impact Assessment

• User Experience: Users cannot abort workflows started via "Run again"
• Workflow Management: Breaks the expected workflow lifecycle
• Reliability: Creates inconsistent behavior that undermines user trust
• Support: Generates support tickets and user confusion