Create an executable migration plan from a legacy warehouse or on-prem data platform to a modern warehouse/lakehouse.
Source and target platforms: [systems]
Data volume, growth, and retention: [TB/PB]
Workloads: [ETL, BI, ad hoc, ML, extracts]
Object inventory: [tables, views, procedures, jobs, dashboards]
SLA and downtime tolerance: [requirements]
Security and compliance constraints: [details]
Current spend and target outcome: [cost/performance goals]
Team capacity and deadline: [details]
Produce:
1. Discovery inventory and workload classification method
2. Dependency graph and migration-wave strategy
3. Decision matrix: rehost, refactor, replace, retire, or retain
4. Schema, SQL dialect, orchestration, and security translation risks
5. Historical-data transfer and ongoing CDC synchronization design
6. Dual-run parity framework for counts, aggregates, records, latency, and cost
7. Consumer cutover plan with explicit go/no-go gates
8. Rollback strategy that avoids divergent writes
9. Decommission criteria and hidden-dependency detection
10. Timeline, staffing model, risk register, and success metrics
Prioritize a representative vertical slice before bulk migration. Identify workloads whose semantics cannot be validated by row counts alone.