https://github.com/bobthebuidler/eth_retry

Provides a decorator to automatically retry calls that fail due to specific transient errors seen in the web3 ecosystem.
https://github.com/bobthebuidler/eth_retry

eth-brownie ethereum evm python python3 web3 web3py

Last synced: about 2 months ago
JSON representation

Provides a decorator to automatically retry calls that fail due to specific transient errors seen in the web3 ecosystem.

• Host: GitHub
• URL: https://github.com/bobthebuidler/eth_retry
• Owner: BobTheBuidler
• License: mit
• Created: 2022-04-27T22:56:20.000Z (over 2 years ago)
• Default Branch: master
• Last Pushed: 2024-05-23T00:48:23.000Z (7 months ago)
• Last Synced: 2024-05-23T00:54:36.830Z (7 months ago)
• Topics: eth-brownie, ethereum, evm, python, python3, web3, web3py
• Language: Python
• Homepage:
• Size: 56.6 KB
• Stars: 10
• Watchers: 4
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: license

Awesome Lists containing this project

README

        

# eth_retry

> Stop transient errors from wasting your time!

`eth_retry` is a Python library that provides one decorator, `eth_retry.auto_retry`.

`auto_retry` will automatically catch known transient exceptions that are common in the Ethereum/EVM ecosystem and will reattempt to evaluate your decorated function up to `os.environ['MAX_RETRIES']` (default: 10) times.

Supports both synchronous and asynchronous functions.

------------

Covers many common transient errors in the EVM ecosystem, including:

- RPC timeouts

- Block explorer API rate-limiting

- Generic exceptions:

  - ConnectionError

  - HTTPError

  - ReadTimeout

  - TimeoutError

- eth-brownie specific errors:

  - sqlite3.OperationalError: database is locked

## Installation:

`pip install eth_retry`

or

`pip install git+https://github.com/BobTheBuidler/eth_retry.git`

## Usage:

```

import eth_retry

@eth_retry.auto_retry

def some_function_that_errors_sometimes():

    i = 0

    am = 1

    doing = 2

    stuff = 3

    return stuff

error_free_result = some_function_that_errors_sometimes()

```

Between attempts, eth_retry will `time.sleep` for a random period between `os.environ['MIN_SLEEP_TIME']` (default: 10) and `os.environ['MAX_SLEEP_TIME']` (default: 20) seconds. The period is randomized to help prevent repetitive rate-limiting issues with parallelism by staggering the retries.

On the `n`th retry, the sleep period is multiplied by `n` so that the target endpoint can cool off in case of rate-limiting.

After `os.environ['MAX_RETRIES']` failures, eth_retry will raise the exception.

## Environment:

```

# Minimum sleep time in seconds. Integer. Defaults to 10.

MIN_SLEEP_TIME=10

# Maximum sleep time in seconds. Integer. Defaults to 20.

MAX_SLEEP_TIME=20

# Maximum number of times to retry. Integer. Defaults to 10.

MAX_RETRIES=10

```