Usage Tracker - Design an API Rate Limiter with Per-User Quotas

Question

## Problem

Design a usage tracker that enforces per-user API rate limits. Each user has a quota of `max_calls` per `window_seconds` (sliding window).

Implement:

```python
class UsageTracker:
    def __init__(self, max_calls: int, window_seconds: int): ...
    def record_call(self, user_id: str, timestamp: int) -> bool:
        # Returns True if the call is allowed, False if rate-limited
    def usage(self, user_id: str, current_time: int) -> int:
        # Returns number of calls in the current window
```

## Example

```
tracker = UsageTracker(max_calls=3, window_seconds=60)
tracker.record_call("u1", 0)   -> True
tracker.record_call("u1", 30)  -> True
tracker.record_call("u1", 50)  -> True
tracker.record_call("u1", 55)  -> False  (3 calls in [0,60))
tracker.record_call("u1", 70)  -> True   (call at t=0 slid out of window)
```

## Follow-ups

1. How does a sliding window log compare to a fixed window counter and a sliding window counter (token bucket) in terms of memory and accuracy?
2. How would you implement this in a distributed system where multiple API servers share state?
3. What happens to memory usage if there are 10M users each with sparse call patterns?
4. Add burst support: allow up to `2 * max_calls` in any 10-second window.

Usage Tracker - Design an API Rate Limiter with Per-User Quotas

Question Details

Full Details

Topics

More from Amazon