Assessment¶
Assessment is how you understand what you are migrating before you commit to conversion and cutover. Strong assessments inventory objects and dependencies, highlight complexity and risk, and support a realistic migration sequence. In this documentation, assessment shows up in three main places: AI assessment through Cortex Code and SnowConvert, the assessment capability inside the Snowflake Migration Plugin skill, and Snowpark Migration Accelerator (SMA) assessment for Spark workloads.
AI assessment (Cortex Code CLI)¶
The AI assessment topic walks you through using the Cortex Code CLI agent with the assessment skill from the Snowflake Migration Plugin. The agent runs a structured, multi-dimensional analysis of source database code and produces a detailed HTML report you can use to build a dependency-aware migration plan and spot likely blockers.
Note
The assessment skill is currently supported for SQL Server and Redshift source platforms.
What the skill helps with
- Categorizing source database objects
- Evaluating conversion complexity
- Proposing a logical migration sequence
Prerequisites (high level)
- Cortex Code CLI installed and configured. See Cortex Code CLI.
- Migration Plugin available in Cortex Code (you are prompted to install it the first time you use the
assessmentskill).
Typical flow
- Start Cortex Code CLI.
- Ask the agent to run the
assessmentskill and supply paths for SnowConvert AI assessment inputs and for the unified report output. - Grant file access when prompted (you can choose to allow reads for future runs).
- Open the generated report at the path you specified.
What you get in the unified report (multi_report.html)
The interactive report supports filtering and CSV export. Its sections align with how you manage scope, sequencing, and risk:
| Section | Purpose |
|---|---|
| Overview | Summary of workload inventory, anticipated manual effort, and entry points into the detailed reports. |
| Exclusion report | Objects that may be excluded after SME review (temporary or staging objects, deprecated or legacy assets, test artifacts, duplicates) so you do not migrate unnecessary scope. |
| Dynamic SQL report | Files containing dynamic SQL, with complexity signals (SQL Server migrations only). |
| Waves report | Dependency-aware waves so prerequisites deploy before dependents (default wave sizing applies; you can ask the agent to customize wave size or order). |
| SSIS package analysis | For SSIS workloads: AI summary (approach, risks, consolidation), metrics, package classification and complexity, and conversion breakdown between automated and manual work. |
For sample prompts (scoped assessments, SSIS-only runs), billing, and legal notices for Cortex and Cortex Code, use the full AI assessment page.
Assessment inside the migration skill¶
When you work through the Snowflake Migration Plugin skill, the assessment stage generates a single self-contained multi-tab HTML report. You can run the analyses individually or together:
| Analysis | What it does |
|---|---|
| Deployment waves | Uses object dependencies to order deployment so dependents come after prerequisites. |
| Object exclusion | Surfaces temporary, staging, deprecated, and test artifacts so you can shrink scope before deployment. |
| Dynamic SQL analysis | Classifies dynamic SQL patterns and highlights native Snowflake coverage vs manual rewrite vs higher complexity. |
| ETL or SSIS assessment | Classifies SSIS packages (ingestion, transformation, export, orchestration, hybrid), maps control and data flow, and estimates migration effort. |
Example prompts such as "Run a full assessment" and "Generate deployment waves" are listed under Assessment in the skill topic.
Snowpark Migration Accelerator (SMA) assessment¶
For Spark API workloads moving toward Snowpark, SMA starts with an assessment that goes beyond simple pattern matching: it builds a semantic model of your codebase. See How the assessment works.
Outputs include
- A readiness score (see the glossary)
- A full list of Spark API references and how they align with Snowpark
- Third-party library imports and compatibility context
- A summary report pulling the inventories together
Generated files from assessment mode are described under Output reports.
The SMA Assessment landing page lists step-by-step and interpretation topics:
- How the assessment works
- Getting started with assessment
- Understanding your assessment results
- Understanding readiness scores
Start from Snowpark Migration Accelerator: Assessment for the full outline.
Choose your path¶
| If you are assessing… | Start here |
|---|---|
| SQL Server or Redshift database SQL (and optional SSIS) with Cortex Code | AI assessment |
| The same workloads through the guided migration skill | Snowflake Migration Plugin skill (assessment stage) |
| Spark API code for Snowpark | SMA assessment |