Design an API for pay computation with retries

You’ve implemented logic that computes a dasher’s pay from order event logs (events include timestamps and state transitions).

Design an HTTP API to expose this as a service. Cover:

• Endpoints (e.g., compute pay for a dasher over a time window; submit events; query results)
• Request/response schema (including datetime formats)
• Idempotency strategy (to safely handle client retries)
• Failure handling and retry semantics (client vs server retries, timeouts, backoff)
• Validation and error codes
• Observability (logging/metrics/tracing)

Assume this API will be called by other internal services and must be robust to duplicate requests and partial failures.