Thread-Safe Token Bucket | Dataford Interview Questions

Implement a thread-safe rate limiter or token bucket algorithm for managing backend LLM API call quotas. Implement run_rate_limiter(capacity, refill_rate, request_times), returning a Boolean for each request timestamp. The bucket starts full, refills continuously at refill_rate tokens per second, and each accepted request consumes one token. Protect shared state so concurrent calls cannot oversubscribe the bucket. Request timestamps are nondecreasing monotonic seconds.

Constraints

1 <= capacity <= 10^6
0 < refill_rate <= 10^6
1 <= request_times.length <= 10^4
request_times contains nondecreasing monotonic timestamps in seconds
Each request consumes exactly one token
The bucket initially contains capacity tokens

Function Signature

def run_rate_limiter(capacity, refill_rate, request_times):

Your question is Thread-Safe Token Bucket. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Constraints

1 <= capacity <= 10^6
0 < refill_rate <= 10^6
1 <= request_times.length <= 10^4
request_times contains nondecreasing monotonic timestamps in seconds
Each request consumes exactly one token
The bucket initially contains capacity tokens

Function Signature

def run_rate_limiter(capacity, refill_rate, request_times):

Your question is Thread-Safe Token Bucket. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.