Purpose: Upload technical files, generate AI-assisted source-to-target lineage, monitor the run, and download audit-ready JSON outputs from your own Azure environment.
A. Before First Use (Administrator – One Time)
Application URL
Open the AZ Lineage URL provided after Azure deployment.
Operator key
Use the operator access key created during deployment. Keep it in a password manager.
Azure OpenAI
An Azure OpenAI endpoint and at least one model deployment must be configured before a real lineage run can complete.
Managed identity
The AZ Lineage managed identity must have the Cognitive Services OpenAI User role on the Azure OpenAI resource.
Test file
Start with one small, known SQL or SAS file whose expected lineage is understood.
B. Run Lineage (Every Run)
1
Open AZ Lineage Open the application URL in a browser. Enter the operator access key and select Connect. Note: Do not share the operator key by email or chat.
2
Check System Status Open Administration & Support. Confirm Blob Input, Blob Output, and Metadata are Healthy; Service Bus and Worker are Configured; and Azure OpenAI is Configured. Note: Select Test Azure OpenAI Connection. Continue only when the test passes.
3
Select Files Choose the technical source files to process. The screen shows the supported formats and upload limits. Start with a small file for the first run.
4
Upload Files Select Upload. Wait until the activity log confirms that the files were uploaded successfully to the customer’s Azure storage.
5
Generate Lineage Select Generate Lineage. Keep the browser open while the run is queued and processed.
6
Monitor the Run Watch Current Run, Activity Log, and Dashboard. Normal status flow is Queued → Running → Completed or Completed with warnings.
7
Review and Download Open the completed run. Review the run summary and warnings, then download lineage_output.json and the supporting output files. Note: AI-generated lineage must be reviewed before regulatory, audit, or production use.
C. What Each Status Means
Status
Meaning / Action
Queued
The run is waiting for the worker. No action is normally required.
Running
Files are being analysed and lineage is being generated.
Completed
The run finished successfully. Review and download outputs.
Completed with warnings
Outputs were created, but some items need review. Check run_summary.json and exception_log.json.
Failed
Open the error details. Fix the stated issue, then use Retry.
Dead-lettered
The run could not be processed after retries. Download the support bundle and contact support.
D. Main Output Files
File
Use
lineage_output.json
Final published source-to-target lineage records.
run_summary.json
Run totals, processing results, and key quality information.
exception_log.json
Warnings, incomplete items, and non-blocking issues.
suppressed_lineage_records.json
Duplicate or unsafe records kept for audit, but not published as final lineage.
cost_token_summary.json
Customer-local token and estimated-cost summary when available; not the Microsoft invoice.
support bundle
Sanitized diagnostic information for support. It is designed to exclude source files, credentials, and final lineage output.
E. Simple Troubleshooting
Problem
Code / Status
What to Do
Cannot connect
Access denied
Re-enter the correct operator key. Do not add spaces before or after it.
AI is not configured
AZL-AI-001
Ask the Azure administrator to configure the Azure OpenAI endpoint and deployment names, then run the connection test.
AI authentication failed
AZL-AI-401 / 403
Ask the administrator to verify managed-identity RBAC on the Azure OpenAI resource.
Deployment not found
AZL-AI-404
Check the Azure OpenAI endpoint and exact deployment name.
Quota or rate limit
AZL-AI-429
Wait and retry, or ask the administrator to check Azure OpenAI quota.
Run failed
Failed
Open diagnostics, read the suggested resolution, fix the issue, and select Retry.
F. Security and Support
Do not send source files, credentials, operator keys, Azure keys, or confidential lineage output by email.
Use the built-in sanitized support bundle when support information is required, and review it before sharing.
Customer files and outputs remain in the customer-controlled Azure environment.
AZ Lineage is AI-assisted. A qualified user should review the output before formal governance, audit, regulatory, or production use.
☐ Connected with operator key ☐ Storage and metadata show Healthy ☐ Queue and worker show Configured ☐ Azure OpenAI connection test passed ☐ Files uploaded successfully ☐ Run completed or completed with warnings ☐ Outputs reviewed and downloaded
Reviewer Notes
Record important warnings, missing lineage, corrections, or approval comments here.
A run is complete only after the outputs have been reviewed and saved securely.
Terms and Conditions
By deploying, accessing, or using AZ Lineage Managed Application, the customer agrees that the product is provided by Global Research Forum as an AI-assisted technical lineage generation tool for processing customer-uploaded SQL, SAS, ETL, metadata, mapping, and similar technical files inside the customer’s Azure environment, and that all uploaded files, Azure configuration, Azure OpenAI endpoint/deployment, permissions, data security, access controls, validation, compliance, and Azure consumption costs remain the customer’s responsibility. The customer confirms that they have the legal right to upload and process all files submitted to the product and agrees not to use the product for unlawful, unauthorized, misleading, harmful, or non-compliant purposes. Outputs generated by AZ Lineage, including lineage JSON, summaries, audit logs, exception logs, and related evidence files, are AI-assisted and may contain errors, omissions, incomplete mappings, unsupported assumptions, or inaccurate interpretations; therefore, all outputs must be independently reviewed and validated by qualified customer personnel before being used for regulatory, audit, compliance, legal, migration, governance, reporting, or business decisions. AZ Lineage does not provide legal, regulatory, audit, financial, security, or compliance advice and is provided on an “as is” and “as available” basis without warranties of accuracy, completeness, fitness for purpose, uninterrupted operation, security, or error-free performance. Global Research Forum is not responsible for customer data, customer Azure configuration, Microsoft Azure or Azure OpenAI availability, third-party service limitations, token usage, cloud charges, customer reliance on unvalidated outputs, or any indirect, incidental, consequential, regulatory, business, reputational, financial, or data-related loss. To the maximum extent permitted by law, Global Research Forum’s total liability for any claim related to AZ Lineage shall be limited to the amount actually paid to Global Research Forum for the product during the three months before the claim, or INR 10,000, whichever is lower. The customer agrees to indemnify and hold harmless Global Research Forum from claims arising from customer-uploaded data, unauthorized use, misuse, incorrect configuration, violation of law or third-party rights, or reliance on unvalidated outputs. These terms are governed by the laws of India unless mandatory applicable law or Microsoft Azure Marketplace terms require otherwise. For support or questions, contact admin@globalresearchforum.in.