obelisk.strategies.retry.ExponentialBackoffStrategy

class obelisk.strategies.retry.ExponentialBackoffStrategy(max_retries: int = 5, backoff: timedelta = datetime.timedelta(seconds=2), max_backoff: timedelta = datetime.timedelta(days=1))

Bases: RetryStrategy

Retry strategy implementing the exponential backoff algorithm. Every failure up to max_retries will result in a sleep of t ** n seconds, with t being the backoff and n amount of failures.

Note that backoff values less than one second will count as zero.

Methods

make()

Create an instance of a RetryEvaluator.

__init__(max_retries: int = 5, backoff: timedelta = datetime.timedelta(seconds=2), max_backoff: timedelta = datetime.timedelta(days=1)) → None

Methods

__init__([max_retries, backoff, max_backoff])
make()	Create an instance of a RetryEvaluator.

Attributes

max_retries
backoff
max_backoff

make() → RetryEvaluator

Create an instance of a RetryEvaluator. Custom strategies should also subclass RetryEvaluator