Handle invalid input at system level
Overview
This question evaluates a candidate's competency in designing end-to-end validation, error handling, and operability for a REST API, covering input schema validation, authentication and authorization, rate limiting, idempotency, error classification, partial-failure strategies, timeouts/retries, observability, and graceful degradation in the System Design domain. It is commonly asked to assess how candidates reason about building robust, production-grade services that handle malformed or oversized payloads and partial failures, and it tests practical application of system-level design principles rather than low-level algorithmic implementation.
Design: Validation and Error Handling for a Top-K Frequency API
Context
You are designing a REST endpoint that computes the top-K most frequent values from a client-provided array. The focus is on end-to-end validation, robustness, and operability rather than the counting algorithm itself.
Assume the endpoint is:
- Method: POST
- Path: /v1/topk
- Purpose: Given an array of primitive values and an integer K, return the K most frequent values with counts.
- Values may be strings or numbers; response ordering must be deterministic (by count desc, then value lexicographically).
Requirements
Design the validation and error-handling strategy covering:
- Input schema validation (types, ranges, size limits, content-type, compression)
- Authentication and authorization
- Rate limiting and quotas
- Idempotency for safe retries
- Consistent error codes and messages (4xx vs 5xx)
- Partial failures (e.g., invalid elements) and response strategy
- Timeouts and retries (client and server), backoff guidance
- Observability (structured logs, metrics, tracing)
- Graceful degradation for oversized or malformed payloads (e.g., sampling or approximate results)
- Provide example API contracts and sample error responses
Solution
Login required