tom.beeckman

Monitoring webapplications with correlation

Discussion created by tom.beeckman on Sep 22, 2009
Latest reply on Sep 24, 2009 by nrtimm
Hi everybody,

I am quite new to Nimsoft monitoring. We started out with end-user-experience monitoring via e2e_appmon probe, but now we are looking to replace our existing monitoring B*C with Nimsoft.

I have 1 thing where I struggle a bit to set it up.

We are currently monitoring webapplications by checking a healthpage -> if the healthpage returns SUCCEEDED, all is ok, if not, the application is down. This is all setup in an loadbalanced environment, where every application runs at least on 2 servers (sometimes more than 2 servers).

Currently we are used to this kind of monitoring:
APP1 runs on SERVER1 & SERVER2

Event: APP1 down on SERVER1
Action: send mail to inform sysadmin, not crictical because APP1 is still running on SERVER2.

Event APP1 down on SERVER1 and SERVER2
Action: send sms to guard (APP DOWN) + email because the application is dead on all servers.

Event APP1 back up on SERVER1 or SERVER2
Action: send sms to guard (APP UP) + email because the application is back up and online.

Event APP1 backup on SERVER1 and SERVER2
Action: send mail to inform sysadmin, not crictical because APP1 is still running on SERVER2.

In short: if APP completely down: critical (sms) if not (email).

Does anyone has experience how to set this up?
I was thinking to use triggers & profiles in the NAS.. but we have like 50 applications with this setup... this will cause a lot of profiles & triggers....I was hoping to implement something more generic...Maybe through LUA?

Any help, suggestions, comments are appreciated.

Best regards
T. Beeckman
Securex Belgium

Outcomes