2👍
✅
If you are concerned about I/O latency (a valid concern), you could log from the web application to a queue using QueueHandler / QueueListener as described in the docs here. The QueueListener can operate on a separate thread and prevent your request handler threads from getting bogged down in logging I/O. If you are using Python 2.x, then QueueHandler / QueueListener are available in the logutils package on PyPI.
1👍
You should log all the data that can be usefull for any future analysis. But you should concern about logging overhead if you are logging too much of data in sync. There can be many options for async logging.
- We are using redis server pubsub model for this ( for example visit this link ).
- The other option could be using django-celery.
- [Django]-Django – why can't I access dynamically-generated attributes in classes in my models.py from admin.py?
- [Django]-Which Server for Django
- [Django]-Placing a <select> field using Django templates with custom bootstrap styles
Source:stackexchange.com