Python Library for Backoff/Retry Strategies
Branch | Unit Tests |
---|---|
latest | |
v. 1.0 | |
develop |
Backoff-Utils is a Python library that provides Python functions and decorators that apply various backoff / retry strategies to your Python function and method calls.
The library has a consistent syntax for easy use, and has been tested on Python 2.7, 3.4, 3.5, 3.6, 3.7 and 3.8.
COMPLETE DOCUMENTATION ON READTHEDOCS: http://backoff-utils.readthedocs.io/en/latest
To install Backoff-Utils, just execute:
$ pip install backoff-utils
Once installed, to import Backoff-Utils into your project you can use:
#: Import the backoff() function.
from backoff_utils import backoff
#: Import the @apply_backoff() decorator.
from backoff_utils import apply_backoff
#: Import backoff strategies.
from backoff_utils import strategies
By design, Backoff-Utils are designed to rely on minimal dependencies. The only dependency they have outside of the Python standard library is:
validator-collection which provides for robust validation functionality.
This library in turn has one external dependency when installed under Python 2.7:
- regex which is a drop-in replacement for
Python's (buggy) standard
re
module.
- regex which is a drop-in replacement for
Python's (buggy) standard
from backoff_utils import strategies
# Using a Function Call
from backoff_utils import backoff
def some_function(arg1, arg2, kwarg1 = None):
# your code goes here
pass
result = backoff(some_function,
args = ['value1', 'value2'],
kwargs = { 'kwarg1': 'value3' },
max_tries = 3,
max_delay = 3600,
strategy = strategies.Exponential)
# Using a Decorator
from backoff_utils import backoff
@apply_backoff(strategy = strategies.Exponential, max_tries = 3, max_delay = 3600)
def some_decorated_function(arg1, arg2, kwarg1 = None):
# your code goes here
pass
result = some_decorated_function('value1', 'value2', kwarg1 = 'value3')
Because now and again, stuff breaks.
Often, when making external API calls to third-party systems, something goes wrong. The internet might glitch. The API we're calling might timeout. Gremlins might eat your packets. Any number of things can go wrong, and Murphy's law tells us that they will.
Which is why we need backoff strategies. Basically, these are techniques that we can use to retry function calls after a given delay - and keep retrying them until either the function call works, or until we've tried so many times that we just give up and handle the error.
This library is meant to be an incredibly simple utility that provides a number of easy-to-use backoff strategies. Its core API is to expose:
- the
backoff()
function, which lets you apply a given backoff strategy to any Python function call, and;- the
@apply_backoff()
decorator, which lets you decorate any function or method call so that a given backoff strategy is always applied when the decorated function/method is called.
The library supports five of the most-common backoff strategies that we've come across:
- Exponential
- Fibonacci
- Fixed
- Linear
- Polynomial
In addtion, you can also create your own custom strategies as well.
For more information about the backoff strategies supported, please see: Strategies Explained
In addition to the basic strategies, the library also supports:
- random jitter
- argument-adjustment on retry
- selective exception capture
- chained backoff strategies
- failure handlers
- success handlers
- cut-off after max delay
- cut-off after max tries
- scaling
- minimum delay
For more information about the backoff strategies supported, please see: Using the Library
We're happy to maintain this library going forward, and would always love to hear users' feedback - especially if you're running into issues.
Please report issues or questions on the project's Github page
For more information on contributing to the Backoff-Utils library, please see: Contributor Guide