It's the first time I post here and I'm a French guy (sorry in advance for my English)
I did several tests with Splunk Light, but I cannot find a Python method to analyze only the incoming logs at the moment I launch the script.
I want to have an infinite loop that will iterate over all the incoming logs.
Do you have a solution to my problem?
It sounds like you dont need a python script and instead you should check out the inputs.conf directive known as "monitor":
http://docs.splunk.com/Documentation/Splunk/6.2.0/Admin/Inputsconf #<-- search for word "monitor"
The following configuration directs Splunk to read all the files in the directory /var/log. It reads them infinitely... as in any new files are automatically read.
The following configuration directs Splunk to read all data in one file within /var/log. Any changes appended to the file are automatically indexed.
I think I need a python script...
I want to analyze all the incoming log to detect certain cases and make a treatment with them (for example send an email, disable an faulty interface,...)
I don't believe you need a python script given what you are trying to do. Splunk monitors every log as it comes in. If you set an alert to look for particular cases and when the case is found you can trigger an action: send an email, run a script to disable the faulty interface etc...