logdissect: CLI utility and Python API for analyzing log files and other data

by do son · Published May 3, 2019 · Updated October 10, 2021

logdissect

Logdissect is a CLI utility and Python library for analyzing log files and other data. It can parse, merge, filter, and export data (to log files, or JSON).

Installing

sudo pip install logdissect

git clone https://github.com/dogoncouch/logdissect.git
cd logdissect
sudo make all

Usage

usage: logdissect.py [-h] [--dhost DHOST] [--grep PATTERN] [--last LAST]

                     [--process PROCESS] [--protocol PROTOCOL] [--range RANGE]

                     [--utc] [--rdhost DHOST] [--rgrep PATTERN]

                     [--rprocess PROCESS] [--rprotocol PROTOCOL]

                     [--rshost SHOST] [--rsource SOURCE] [--shost SHOST]

                     [--source SOURCE] [--linejson LINEJSON] [--outlog OUTLOG]

                     [--label LABEL] [--sojson SOJSON] [--pretty] [--version]

                     [--verbose] [-s] [--list-parsers] [-p PARSER] [-z]

                     [-t TZONE]

                     [file [file ...]]



positional arguments:

  file                  specify input files



optional arguments:

  -h, --help            show this help message and exit

  --version             show program's version number and exit

  --verbose             set verbose terminal output

  -s                    silence terminal output

  --list-parsers        return a list of available parsers

  -p PARSER             select a parser (default: syslog)

  -z, --unzip           include files compressed with gzip

  -t TZONE              specify timezone offset to UTC (e.g. '+0500')



filter options:

  --dhost DHOST         match a destination host

  --grep PATTERN        match a pattern

  --last LAST           match a preceeding time period (e.g. 5m/3h/2d/etc)

  --process PROCESS     match a source process

  --protocol PROTOCOL   match a protocol

  --range RANGE         match a time range (YYYYMMDDhhmm-YYYYMMDDhhmm)

  --utc                 use UTC for range matching

  --rdhost DHOST        filter out a destination host

  --rgrep PATTERN       filter out a pattern

  --rprocess PROCESS    filter out a source process

  --rprotocol PROTOCOL  filter out a protocol

  --rshost SHOST        filter out a source host

  --rsource SOURCE      filter out a log source

  --shost SHOST         match a source host

  --source SOURCE       match a log source



output options:

  --linejson LINEJSON   set the output file for line by line JSON output

  --outlog OUTLOG       set the output file for standard log output

  --label LABEL         set label type for OUTLOG (fname|fpath)

  --sojson SOJSON       set the output file for single object JSON output

  --pretty              use pretty formatting for sojson output

Parsers

The output of –list-parsers option:

==== Available parsing modules: ====



ciscoios        : cisco ios parsing module

emerge          : gentoo emerge log parsing module

linejson        : logdissect object-per-line JSON parsing module

sojson          : logdissect single object JSON parsing module

syslog          : syslog (standard timestamp) parsing module

syslogiso       : syslog (ISO timestamp) parsing module

syslognohost    : syslog (standard timestamp, no host) parsing module

tcpdump         : tcpdump terminal output parsing module

webaccess       : web access log parsing module

windowsrsyslog  : windows rsyslog agent log parsing module

Examples

logdissect --last 10m auth.log

logdissect --last 1h --process systemd --process CRON messages

logdissect -v --range 20160202020202-20170227213200 --label fpath messages debug apache2/error.log

logdissect -s --outlog myaccess.log --grep 192.168.1.56 --last 30d --label fname /var/log/apache2/access.log /var/log/apache2/other_vhosts_access.log

Notes

metadata: logdissect uses file modification times to assign years to syslog date stamps. This allows it to parse logs that span more than one year without a problem. If you are copying log files, always use cp -p (or cp –preserve=timestamps) and scp -p to preserve original mtimes and other file metadata.
Re-parsing: If you are planning on parsing an output file back into logdissect at some point, using JSON is highly recommended. The JSON output modules use date stamps that include a year. Re-parsing a standard log output file will cause problems if the file has a different mtime than the original logs.
–range shortcuts: The range module will fill in your ranges with zeroes if they are shorter than 14 characters. If you want to get a range of 20170204120000 to 20170204130000, you can save time and use 2017020412 and 2017020413.
–last options: The last option should be a number followed by either ‘s’ for seconds, ‘m’ for minutes, ‘h’ for hours, or ‘d’ for days (e.g. –last=20m).
Multiple options: All non-time-based filters can be used more than once.

logdissect 2.0 and above provides a stable API for parsing log lines and files. For more information on library usage, see the API documentation.

Source: https://github.com/dogoncouch/