Extend RedisSettings to include redis Retry Helper settings

[mernmic]

proposed fix for #386

[codecov]

Codecov Report

Merging #387 (e54eace) into main (94cd878) will increase coverage by 0.10%.
Report is 1 commits behind head on main.
The diff coverage is 100.00%.

❗ Current head e54eace differs from pull request most recent head a1ac23a. Consider uploading reports for the commit a1ac23a to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #387      +/-   ##
==========================================
+ Coverage   96.27%   96.38%   +0.10%     
==========================================
  Files          11       11              
  Lines        1074     1078       +4     
  Branches      209      209              
==========================================
+ Hits         1034     1039       +5     
  Misses         19       19              
+ Partials       21       20       -1     

Files	Coverage Δ
arq/connections.py	90.32% <100.00%> (+0.25%)	⬆️

... and 1 file with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8321dc1...a1ac23a. Read the comment docs.

[JonasKs]

The tests has been fixed in #382, so don’t worry about those failing pipelines on some Python versions for now.

As you can see from the codecov bot above, this repo strive to have 100% codecov. Would love if you added tests for these settings.
If you’re unsure how to proceed, please don’t hesitate to ask. 😊

[mernmic]

@JonasKs, added some tests which improved coverage. Still one partial, but might be from testing failures?

[JonasKs]

I'll try to review again this weekend / early next week 🙇‍♂️

[JonasKs]

Looks good to me! 😊

[mernmic]

Looks good to me! 😊

Thanks for taking a look, just pushed a change for the Exception type.

We've been using the above changes in our systems and everything has been working as expected so far.

A heads up I'll be offline from the first few weeks of February, so if anything comes up that's the reason for any radio silence on my end.

Thanks again!

[JonasKs]

Thanks for the PR and for testing in your environment!
Enjoy your offline time☺️

I'd also recommend double checking that maintainers can edit your PR, in case Samuel finds time to review/release while you're gone. 😊

[maximgrigorov]

Hi there,

is it possible to merge this fix to main branch?

[akira]

Hello- thanks for the fix! Any thoughts on merging this? Would be great to have for my use cases! 🙇

[mernmic]

Workaround while waiting for PR merge.

Initialize ArqRedis with a connection pool including your retry helper.

import redis
import arq

class WorkerSettingsBase:
    redis_pool = arq.ArqRedis(
        connection_pool=redis.asyncio.ConnectionPool.from_url(
            url=RedisSecrets.REDIS_URL,
            retry_on_timeout=True,
            retry_on_error=[
                redis.asyncio.TimeoutError,
                redis.asyncio.ConnectionError,
            ],
            retry=redis.asyncio.retry.Retry(
                backoff=CustomRedisBackoff(),
                retries=5,
            )
        )
    ),
   ....

if __name__ == "__main__":
    arq.run_worker(WorkerSettingsBase)

redis_pool parameter can be found in docs here:
https://arq-docs.helpmanual.io/#arq.worker.Worker

Example custom backoff class, we are using for more granular access and logging but you can use any redis retry helper

class CustomRedisBackoff(AbstractBackoff):
    def __init__(self, cap=0.512, base=0.008):
        """
        `cap`: maximum backoff time in seconds
        `base`: base backoff time in seconds
        """
        self.backoff = ExponentialBackoff(cap=cap, base=base)

    def compute(self, failures):
        sleep = self.backoff.compute(failures)
        logger.info(
            {
                "message": "redis connection error, triggering backoff",
                "failures": failures,
                "backoff": repr(self.backoff),
                "sleep (secs)": sleep,
            }
        )
        return sleep

docs on redis retry helpers:
https://redis-py.readthedocs.io/en/stable/retry.html

[samuelcolvin]

Thanks so much for this, sorry for the long delay.

[samuelcolvin]

@JonasKs @Kludex we might be able to use redis's retry logic to remove our retry stuff - https:/samuelcolvin/arq/blob/94cd8782b4f0764a17962186a349d32125cb98e3/arq/connections.py#L258-L286

[samuelcolvin]

v0.26.0b1 is released, please try it, I'll release v0.26 at the end of the week, see #441.