This sounds like an architectural question.
If I understand your requirements properly, what you are looking for is something like this,
- You have a Python based system with one or more standard Python loggers writing to text file(s).
- You need to allow several users to access this logging data in real time, using a variety of filtering techniques to control what they access.
- Based on what they find from analysing this information, they may need to issues commands to the system to modify it's behaviour.
Given this set of requirements, I would suggest the following approach.
- Create a centralized database to store your log records.
- Populate the database using the standard Python logger and a custom logging handler that communicates directly from your logger to the database.
- Develop a server based application that mediates between the web based clients and the database.
- Use a message queue between your server logging manager and the original system to provide a transport mechanism for commands from your users to the system.
There many packages and tools that can be used to realize this scenario but I will assume that you want a primarily Python based solution and make a number of concrete suggestions.
- Use MongoDB as your database.
- Use MongoLog as the logging handler to interface between your original system and the MongoDB database.
- Develop your server application in Python using Twisted as your core framework.
- Use ZeroMQ as your message queue between your logging manager and your original system.
If you are prepared to consider using non-Python based solutions I would suggest using Meteor (a Javascript based environment) to develop both the web client environment and the logging manager.