What is Idempotency?
Idempotency is the property by which repeating an operation with the same identity produces no additional unintended effects after the first success. It is important when a client cannot determine whether an earlier request completed.
How Idempotency works
An idempotent service treats repeated execution of the same logical command as one effect, even when network uncertainty causes the caller to resend it. Servers usually persist a key alongside request identity, processing state, and the prior response or resulting resource. The guarantee has a scope and retention period; it is not an unlimited deduplication promise. In media systems it protects upload creation, encoding submissions, billing actions, and webhook handling around retry-prone boundaries.
Key facts
- 1A robust implementation compares the retried request’s relevant parameters with the stored operation; returning an old result for the same key but different input can silently associate the wrong media.
- 2Concurrent duplicates require an atomic reservation or uniqueness constraint, because a check followed by a separate insert allows two workers to observe absence and perform the effect twice.
- 3HTTP method semantics and application idempotency are separate: PUT is defined as idempotent at the protocol level, while a POST can be made retry-safe through an operation-specific key.
When Idempotency matters
Assign an idempotency key to retried operations in services you build, or when calling third-party APIs that document key support, so the server can return the original result instead of creating another resource. Transloadit’s API does not accept an idempotency key, so deduplicating repeated submissions is the integrating application’s responsibility; any key you use must identify the same logical operation without being reused for unrelated work.