Ad Widget

Collapse

Best practice needed in logs analyser

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • etienno@gmail.com
    Junior Member
    • Aug 2012
    • 2

    #1

    Best practice needed in logs analyser

    Hello all,

    (sorry if this is a duplicate)

    we have started a zabbix monitoring project in our company, and we would like to have some advices in the way we need to get the KPI from our applications.

    First of all we plan to monitor few dozen applications distributed on some 100 servers.

    We have chosen to monitor the KPI in the application logs of each application / server. What is the best approach to do that? I mean what is the best design or architectural approach in the way the scripts should extract the KPI information from the logs?

    One of our question is the best way to scan the logs. For security reason, we have chosen to use a "push pattern" to push the data on the Zabbix server, instead of pulling the data from the Zabbix server.

    For example, we plan to make a kind of "tail -f " on each server on the application logs and then make a "one minute extract" in a "awk" script. The awk script would extract the KPI from the "One minute Extract" and send the "kpi extraction" by sftp/scp to the zabbix server. Is that a valide way to scan the logs, or is there some ready made pattern?

    thanks and regards,

    Etienne
    Montreal
Working...