Ensure prune_logs can be used more efficiently when dealing with resource limits

[alextreme]

We have our prune_logs task here:

https://github.com/maykinmedia/django-log-outgoing-requests/blob/main/log_outgoing_requests/tasks.py

This works if you have unlimited resources, however when running this for a very large database (eg the task hasn't been run before) and in a resource-restricted environment (k8s resource limits) the following can occur:

In [1]: from log_outgoing_requests.models import *                                                                                                                                                                           

In [2]: OutgoingRequestsLog.objects.first()                                                                                                                                                                                  
Out[2]: <OutgoingRequestsLog: api.twitter.com at 2024-02-07 12:10:00.254395+00:00>

In [3]: OutgoingRequestsLog.objects.count()                                                                                                                                                                                  
Out[3]: 697150

In [4]: from datetime import timedelta                                                                                                                            

In [5]: now = timezone.now()                                                                                                                                      

In [6]: settings.LOG_OUTGOING_REQUESTS_MAX_AGE                                                                                                                    
Out[6]: 30

In [7]: OutgoingRequestsLog.objects.filter(timestamp__lt=now - timedelta(180)).delete()                                                                           
Geëlimineerd

I'd recommend to perform this delete in batches (it wasn't a problem when I did this in batches of ~60000 logs)