Successful tasks grow beyond save_limit

[bob-r]

Detected with two simultaneously running clusters. Looks like a race condition in qcluster: https://github.com/Koed00/django-q/blob/master/django_q/cluster.py#L416

[peteyb]

Is it possible to merge this into the next version release. We have used Django-Q on two distinct projects, one works fine but the other does not clear down successful tasks.

Both projects are running python:3.6-stretch, django-q==1.0.1, dango=1.11.x and deployed to identical production environments

The only differences I can see if the frquency of tasks being processed with the one not working processing many more requests than the one which clears down successful tasks to the limit.

[jur-clerkx]

@Koed00 Would still be really nice to see this implemented. Also running multiple clusters and had to implement a scheduled task to clean up old tasks results because the list keeps growing. Happy to help if needed.