shorten-hadoop-log

#!/usr/bin/env ruby
# -*- encoding: utf-8 -*-
#

#
# log shortener for Hadoop logs
#
# History:
#   2015/12/04,hancheol.cho - Simple rule-based (1:1) log shortening algorithm
#

PATTS = {
  /org\.apache\.hadoop/ => "<OAH>",
  /[Nn]amenode/ => "<NN>",
  /[Rr]esource[Mn]anager/ => "<RM>",
  /[Dd]ata[Nn]ode/ => "<DN>",
  /[Nn]ode[Mm]anager/ => "<NM>",
  /[Cc]ontainer[Mm]anager/ => "<CM>",
  /LinuxContainerExecutor/ => "<LCE>",
  /[Aa]uthorization/ => "<AuthZ>",
  /[Mm]anager/ => "<Mgr>",
  /[Ss]uccessful/ => "<OK>",

  /KERBEROS/ => "<KRB>"
}



$stdin.each do |line|
  line = line.rstrip
  dt, tm, logLvl, logMsg = line.split(" ", 4)

  PATTS.each do |pat, val|
    logMsg = logMsg.gsub(pat, val)
  end
  puts "#{dt} #{tm} #{logLvl} #{logMsg}"
end