Which rate-limiting algorithm lets a client burst up to N requests at once, then continue at a steady average rate?

Options

• ✓ Token bucket Correct
• ✗ Leaky bucket that queues requests and drains them at a fixed rate
• ✗ Fixed window counter
• ✗ Sliding log

Why

Tokens refill at rate R up to a capacity of N; each request spends one. A client that was idle has a full bucket and may burst, which suits real traffic. A leaky bucket instead smooths output to a constant rate.

Review this in your daily deck → All cards in Resilience