Developing for Splunk Enterprise

File input buffer size vs reading file in with modular input

New Member

I have a set of files that need some processing before indexing with Splunk. The multi line events can be broken out by the "Date\/Time:" regex. I was playing around with python modular input and used the following code to read the file in:

with open('text.log', 'r') as f:
    for line in f:
        event = Event()                
        event.stanza = input_name = line
        event.sourceType = 'mylog_event'
        event.index = 'mylog_index'


category = Custom
description = My Log Events
pulldown_type = 1
disabled = false

When I use the modular input, the events are not broken out correctly and they are all single line each. Looking at the source of the event, each source is 10 lines. With the same props.conf file when I upload the file using the default file input, the events are broken out correctly and when looking at the sources they are all a multiple of the lines in each event.

How do I check how the default file input complete props.conf is?
What changes the number of lines grouped into a source event when reading from a file?


0 Karma

New Member

After searching and some trial and error I found the btool command that seems to show the layering for each sourcetype and their full props.conf properties. I was using this with Splunk Enterprise 7.0.

  1. $SPLUNK_HOME/bin/splunk cmd btools props list --debug This will list out the full props.conf for all supported sourcetype. The --debug option will show which props.conf file contributed to each property.
  2. $SPLUNK_HOME/bin/splunk cmd btools props list list --debug This will list the full props.conf used for the sourcetype that the pattern matches.
  3. $SPLUNK_HOME/bin/splunk cmd btools --app= props list list --debug Specify the specific app context to list the props.conf for the matching sourcetype.
0 Karma