resonate-external-system-of-record-pattern-typescript

To maintain cross-system data consistency without distributed transactions, stage writes to a reference system first, then commit to the system of record to enforce the Write Last, Read First principle.

The Write Last, Read First pattern orders operations by staging writes to a reference system before committing them to the system of record, ensuring system-wide consistency without transactions.

Handle crash recovery and idempotent retries in multi-system integrations by using durable execution to automatically recover operations and tolerate crashes without duplicating writes to the system of record.

Use invariant checks with ctx.panic for external system conflicts when coordinating writes between systems and you need to immediately halt execution upon detecting data consistency violations.

Yes, this durable execution pattern is implemented for TypeScript integrations, providing crash-recovery guarantees and idempotent retries for coordinating writes across external systems of record.

Coordinating writes without distributed transactions requires careful staging in a reference system and invariant checks to handle conflicts, meaning you rely on durable execution rather than ACID guarantees across systems.