Do not start any new threads during shutdown

[tor-gu]

This is a fix for #193

I was only able to reproduce the issue in an AWS Glue container. Here is a minimal example.

# gluejob.py
import watchtower
import logging
from awsglue.context import GlueContext
from pyspark.context import SparkContext


if __name__ == '__main__':
    # Set up just enough of Glue to trigger the issue
    sc = SparkContext()
    glueContext = GlueContext(sc)

    # Add a watchtower handler to the root logger
    # Make sure the log_stream depends on the logger_name
    logging.basicConfig(level="INFO")
    logger = logging.getLogger()
    handler = watchtower.CloudWatchLogHandler(
        log_group_name="/foo", 
        log_stream_name="stream_{logger_name}"  # the default log_stream would also work
    )
    logger.addHandler(handler)

    logger.info("Exiting...")

SCRIPT_FILE_NAME=gluejob.py

# Install watchtower and submit the glue job
docker run -it --rm \
    -v ~/.aws:/home/hadoop/.aws \
    -v $PWD:/home/hadoop/workspace/ \
    -e AWS_PROFILE=$AWS_PROFILE \
    --name glue_watchtower \
    public.ecr.aws/glue/aws-glue-libs:5 \
    -c "pip3.11 install watchtower==3.4.0; spark-submit /home/hadoop/workspace/$SCRIPT_FILE_NAME"

On my machine, it hangs with this message most (but not every!) time. There are evidently some race conditions as things are deallocated.

INFO:py4j.clientserver:Closing down clientserver connection