A distributed workflow needs a localized repair without weakening observability or rollback safety. The affected area is distributed cache.
Repair the distributed cache command handling implementation so it accepts first operations, recognizes safe repeats, and rejects key reuse with a different payload.
Evidence
- Affected surface: Cloud-native full-stack / distributed cache.
- Observed failure family: enforce idempotent operations.
- Scope policy: retained must equal true.