Quick Answer
Rate limiting at service level: token bucket or fixed window algorithms in middleware. At API gateway: centralized rate limiting with shared state in Redis. Throttling is softer - slow down requests instead of rejecting. Per-user, per-IP, or per-API-key limits. Return 429 with Retry-After header. Use sliding window for smoother limits without burst at window boundaries.
Answer
Protect services from overload using rate limits. Can be implemented at API Gateway or per-service level. Patterns: Token bucket, leaky bucket.
S
SugharaIQ Editorial Team
Verified Answer
This answer has been peer-reviewed by industry experts holding senior engineering roles to ensure technical accuracy and relevance for modern interview standards.
