Retry a request with backoff

A request that fails once will often succeed on the next try. Retrying immediately just hammers a struggling server, so each wait should be longer than the last.

Implement retry(task, { attempts, baseDelay }) where task returns a promise. Retry on rejection until attempts run out, doubling the delay each time.

What it has to do

• Resolve as soon as any attempt succeeds.
• Reject with the last error once attempts are exhausted.
• Double the delay after each failure, starting at baseDelay.
• Make no extra call after the final attempt.

Your workspace