Backend / Reliability
API retry & idempotency lab
Simulate an operation that commits after a client timeout, then change retry counts and deduplication to compare server-side side effects.
Scenario controls
Assumption: every attempt is received and the server commits even after the client timeout. Deduplication is atomic and shared across attempts.
Server writes
3
Client outcome
Timed out
Repeated attempts generated duplicate effects.
Attempt trace
- Attempt 1client timed out
Write executed. The server commits the operation even if the client stops waiting.
- Attempt 2client timed out
Write executed. The server commits the operation even if the client stops waiting.
- Attempt 3client timed out
Write executed. The server commits the operation even if the client stops waiting.
What this model does — and does not do
This is a deterministic teaching model with synthetic inputs. No real credentials, traffic, signing keys, attestation services, authentication endpoints, or customer data are processed. A production system needs actual cryptographic checks, authoritative persistence, concurrency control, and integration tests.
Inspect the actual lab model
Download the deterministic TypeScript model and its automated regression tests as a self-contained source bundle. It contains no production app code, real secrets, or platform attestation implementation.
Download reference source + tests