Rq retry

[amks1]

Added support for the Retry() class of RQ. Tested working.

[liqb]

@amks1 It's great that you have retry added. I tried the retry branch and found that it does not support try with interval argument.

retry=Retry(max=3, interval=10)

I checked rq docs and found that --with-scheduler parameter should be used when start a worker, so I suggest you add it to worker command. Thanks.

[mllife]

Not working for me with redis. I added retry=Retry(max=3, interval=[1, 1, 1]) to enqueue method

My 'on_failure' method is something like this; on_failure=report_failure,

def report_failure(job, connection, type, value, traceback):
    # 'value' is the error msg. thrown by the error function
    logger.info("entered report_failure")
    logger.info(f"{job.retries_left=}")
    logger.info(f"{value=}")
    if job.retries_left:
        # redo
        logger.info("inside if")
        logger.info(
            f"{job.retries_left} retries left. Debug msg. '{value}'",
            extra={"clp_fields": {
                "chainId": job.id,
                "meta": job.meta,
                "process": "worker"
            }},
        )
        return
    # last try
    logger.info("outside if")
    logger.info(f"{job.args[0]['input']=}")
    # log final error
    logger.error(
        f"All retries failed. Error msg. '{value}'",
        extra={"clp_fields": {
            "chainId": job.id,
            "meta": job.meta,
            "process": "worker",
        }},
    )

The code only enters the report_failure 1 time and then don't do anything after that, it is supposed to call the retry worker method automatically. This was working with python rq library.