AI assessment¶

This topic helps you use the Cortex Code CLI agent to assess the source code to be migrated to the target Snowflake database.

Refer to Cortex Code CLI for more information on how to use the Cortex Code CLI agent.

The Cortex Code CLI agent uses the snowconvert-assessment skill to analyze the source database code. It invokes a structured, multi-dimensional analysis of the source database code and generates a detailed actionable report. The generated report helps create a clear, dependency-aware migration plan by identifying potential road blocks and risks associated with the migration.

Core functionality of AI assessment¶

You can perform AI assessment on source database code by invoking the snowconvert-assessment skill of the Cortex Code CLI agent. This skill enables the Cortex Code CLI agent to perform assessments by:

  • Categorizing the source database objects

  • Evaluating the conversion complexity of the source database code

  • Establishing a logical migration sequence

The agent then generates a unified comprehensive report in an html format at the specified file location.

Prerequisites to run AI assessment¶

  • Install Cortex Code CLI. Refer to Cortex Code CLI for more information on how to install the Cortex Code CLI agent.

  • Reports generated by SnowConvert AI during a prior run of the same migration project.

Steps to run AI assessment¶

Follow these steps to run AI assessment using the snowconvert-assessment skill of the Cortex Code CLI agent. A comprehensive unified report will be generated in the file location specified in the prompt.

  1. Launch Cortex Code CLI in a terminal window.

  2. Prompt the agent to use the snowconvert-assessment skill. The prompt should also include the file location for the reports generated by SnowConvert AI and the file location for the unified report to be generated.

    Example prompt

    Use skill snowconvert-assessment, create a comprehensive assessment report for <filepath location of SCAI reports>, and put the results in <filepath location of the assessment report>.
    Copy

  3. You may encounter questions asking for read access to files. Selecting always allow this operation in the future can prevent these questions from appearing in future prompts.

  4. Review the unified report generated at the file location specified in the prompt.

Output of AI assessment¶

On completion of the AI assessment, you will find a comprehensive analysis report multi_report.html located in the file location you specified in the prompt. This is an interactive report that allows you to view objects and apply filters based on the object status. You can export the object lists into csv files. The four sections of the unified report consist of:

Each section of this report covers the following critical aspects of risk management during the migration process.

  1. Identifying migration scope: The Exclusion report helps identify deprecated or unused objects that would have no impact if they were not migrated, preventing unnecessary migration efforts.

  2. Targeting code conversion pain points: The Dynamic SQL report helps target SQL code containing dynamic SQL statements, which are often the most complex to convert and need to be evaluated for complexity.

  3. Establishing a migration sequence: The Waves report helps organize the objects into logical execution groups called “waves”, ensuring that all prerequisites for Wave Two are fully converted and deployed during Wave One. This structured approach guarantees “bottom-up” deployment, eliminating the risk of “missing dependency” errors that can stall complex migration projects.

  4. Replicating the exact workflow of SSIS packages: The SSIS package analysis report helps classify SSIS packages (for SQL Server Integration Services) into the categories of ingestion, transformation, and configuration. It also includes an assessment of the complexity of the package, and Directed Acyclic Graphs (DAGs) reflecting workflows inside the packages.

Overview¶

This section contains a summarized view of the assessment results for the migration workload. It includes the total workload inventory, anticipated manual effort, and quick access tiles to the detailed exclusion report, dynamic SQL report, waves report and SSIS report. Select any tile to view the detailed reports. These reports can also be accessed from the left navigation bar.

Exclusion report¶

This section contains a list of objects that can be potentially excluded from migration, based on additional review. The Cortex Code CLI agent intelligently flags deprecated files, temporary staging objects, test objects, and duplicate objects found in source database code. These objects can be excluded after further review with the subject matter experts (SMEs) before the code conversion kickoff.

Under the Detailed Objects Analysis section, you will find a list of objects that are flagged as:

  • Temporary/Staging

  • Deprecated/Legacy

  • Testing objects

  • Duplicate objects

Excluding these objects from the migration may help in reducing the migration effort and time.

Dynamic SQL report¶

This section contains a list of all source code files that were found to contain dynamic SQL. Select any file to view the dynamic SQL code. The detailed view shows the sections of the SQL file containing dynamic SQL and the corresponding line numbers. Select Complexity to view the assessed migration complexity of the dynamic SQL code in the file.

Note

This section is available for migration from SQL Server databases only.

Waves report¶

By default, each wave contains 40 objects. The Cortex Code CLI agent analyzes the relationships between the code objects and ensures that all prerequisites for Wave Two are fully converted and deployed during Wave One. This structured approach guarantees “bottom-up” deployment, eliminating the risk of “missing dependency” errors.

You can prompt the Cortex Code CLI agent to customize the waves report by specifying the number of objects per wave, or changing the order of migration to align with business needs.

SSIS report¶

The SSIS Report helps determine the feasibility and effort required to migrate data workloads consisting of SQL Server Integration Services (SSIS) packages to Snowflake. The core components of this report are:

  1. AI summary: This section contains an overview of the migration readiness of the SSIS packages analyzed by the Cortex Code CLI agent. It consists of migration workload overview, source and destination data flows, connection managers, and an overview of tasks that contribute to the complexity of migration. The AI summary section is further subdivided into three parts:

    • Recommended Migration Approach containing details on methods of migration and the number of packages it can be applied to.

    • Key risks containing details on connectivity gaps, process dependencies on external services like email-based data delivery, manual scripting effort required rewrite custom scripts that cannot be converted automatically, compliance requirements (such as HIPAA, FERPA) that may require data masking and role-based access control (RBAC) implementation.

    • Consolidation opportunities containing suggestions for applying the same migration approach to similar packages.

  2. Key metrics: This section contains a quantified technical footprint of the source SSIS environment. The total component counts, number of transformation pipelines required for data flows, number of workflow and orchestration tasks for control flows and the number of databases with corresponding file connections are all summarized here.

  3. Package classification: This section contains charts to show the classification of package categories and complexity. The interactive table contains a complete list of packages sorted by package name. Enter the package name to search on a specific package. Select options from the classification and complexity dropdowns to view a filtered list of packages. Select the package name in the list to view the complete AI analysis for the package. Select Control Flow DAG to open the Directed Acyclic Graph (DAG) depicting the workflow inside the package.

  4. Component conversion breakdown: This section shows the number of SSIS package compnents that can be automatically converted to Snowflake and the number of components that require manual intervention.

Sample prompts¶

Use the following sample prompts to customize your assessment.

The Cortex Code CLI agent can be scoped to target specific assessments, such as exclusion candidates, dynamic SQL objects, wave reports, or SSIS packages.

To run selected assessment sections

Create an assessment for [wave report|dynamic SQL|object exclusion|ssis package analysis] using SnowConvert assessment files and generate an html report in <filepath location of the assessment report>.
Copy

To perform SSIS package assessment

Perform SSIS package analysis and assessment for this workload: [PATH]. Place results in [filepath location of the assessment report]. Follow instructions of skill snowconvert-assessment and respect rules. Every workflow step must be followed in the order of execution.
Copy

Billing and cost considerations¶

AI assessment consumes Snowflake credits based on the token consumption for interacting with Snowflake Cortex LLM functions. Refer to the current rates in Snowflake Service Consumption Table.

Language: English