The logging of the logs input is updated to make use of structured
logging. Each logs input instance will create an unique ID, that will be
logged as "input_id".
The input and harvester will also add file state information like:
source, state_id, finished, os_id. The os_id holds the inode, and
state_id is the ID that is used internally by filebeat to track the
file.

By filtering the logs for one or the other ID, one can track all
logs at a different level. Filtering by "state_id" allows us to monitor
all logs and state changes for a given file, even if the file is
renamed, closed and reopened.