1

We are running Cognos 8.2 on a Windows 2003 server. I am currently monitoring that its processes and web page are up. After a night where several reports hung and a few days latter the whole system hung allowing user to log in but not do anything else it has become obvious that the current monitoring is not enough. Are there tools out there that allow you to check the over health of Cognos, report run times, etc.?

Regards, Will

Will
  • 51

3 Answers3

1

Will, Are you also getting the dreaded CAM-AAA-0027 errors?

We're monitoring the Cognos service to verify it is running. Oddly I get more false positives on my monitoring (with BigBrother) where the service shows as "Starting" after it's thrice weekly reboots and everyhting is up and running but the service status never changed to "Started".

We have some reports that run a few times a day that track Cognos usage and response times. They're reporting out of the logging database amongst other places. If the times are out of control then we know about it then. Mostly our solution to the memory leak (which is what causes the slowdowns) has been to reboot the servers three times a week. I think we could probably get by with once a week but the application owners would rather have three known short outages than the dreaded slow down and stop.

Laura Thomas
  • 2,855
1

I would suggest looking at new servers and also possibley looking at monitoring software from IBM Tivoli monitoring software.

Here is a link to the software and it would help maintain and manage your systems health.

http://www-01.ibm.com/software/tivoli/products/monitor/

Its also not very costly and easy to manage use.

0

Cognos includes a package called Audit that links to an internal database with logging and performance info if you want to measure your system's health. To use it, you need to create a report with that package and then you create your own monitoring dashboard with the metrics offered by the Audit package.

To check if it's functional, what we have is a cron job that runs a script that fetches the front page and sends a mail if the HTTP response status is anything other than 200 OK.

RAKK
  • 151