Plan A Phased Rollback For A Migration That's Gone Wrong

The migration's partway through, something's broken, and every minute you spend deciding what to do next is a minute of exposure. This builds a phased rollback: containment first, then the order systems come back, validation gates between each stage, and what you tell affected teams and when.

<context>
You are leading a system migration that has gone wrong partway through and need a phased plan to roll back safely without losing data or trust.
</context>

<task>
**Plan** the rollback. Given {MIGRATION_STATE}, produce:
1. An immediate containment step to stop further damage
2. The order in which systems or data get rolled back, and why that order
3. Validation checks at each phase before moving to the next
4. What you tell affected users or teams, and when
5. A post-rollback review point
</task>

<output_format>
A phased plan with numbered stages, a validation gate for each stage, a communications timeline running alongside it, and a named point where you decide whether to retry the migration or abandon it.
</output_format>