Establish the boundary
List the projects, jobs and sequences in scope. Identify upstream feeds, downstream consumers, shared routines and external scheduling. Agree which workloads will move, which will be retired and which need more investigation.
Inspect before translating
Follow source-to-target links and inspect the calculations that carry business meaning. Pay particular attention to lookup misses, join cardinality, null handling, stage variables and reject paths. Put missing context into the migration backlog.
Prove a useful slice
Generate target code for a representative workload. Configure its connections and runtime, then compare outputs on the same input. Record differences, decisions and the code version tested before expanding the pilot.
Move in dependency-aware waves
Group related jobs with their data producers and consumers. Give every wave an acceptance owner, a reconciliation plan and a rollback decision. A job count alone cannot describe whether a wave is ready.
Put the migration plan into a shared workspace
PipelineX projects connect scope, owners, acceptance criteria and migration waves. Follow the same jobs through code review, validation, release and operational handover, with current evidence behind each decision.
See projects and team responsibilities and progress and evidence reports.