Reference · Glossary
Exponential backoff
Last updated
Retry strategy: **wait longer after each failure** (1s, 2s, 4s…) to avoid hammering a rate-limited API.
When to use
429 rate limits, transient 5xx errors, queue workers calling LLM APIs.
When not to
401 bad credentials or 400 invalid payload — fix the request first.
Example
wait = 1
for attempt in range(5):
try:
return call_api()
except RateLimitError:
time.sleep(wait)
wait *= 2