DB2 UDB Workgroup V8.1/Linux

We want to keep stats on uptime of our db2 system to calculate our availability. (ie. 99.5% uptime, etc)

This has to be a pretty common report in organizations, so I'm curious to hear how people have solved this problem.

There's no problem putting an entry somewhere when the database manager comes up, that's the easy part. The hard part is when the database goes down unprompted, ie. crashes/traps.

Presumably I'll need something outside of DB2 to determine this. I was thinking of either a watchdog process that "pings" db2, or a cron job that starts every minute and inserts (or updates) a timestamp into a table to prove everything is ok.

Thanks in advance for your input.

Thought I would resurrect this one, see if anyone has any ideas.

In one of the applications I support we use Big Brother monitoring tool to control status of a lot of things, including an Oracle db. It's freeware (http://bb4.com/).

Basically, it is a collection of scripts that run periodically and then report their respective status on a web page (potentially sending an email alert, too), where you can drill down and view status change history. For Oracle I have a java app. "plug-in" because I need some tricky reporting but if you only want a "ping" to the database then a small script querying sysibm.sysdummy1 would be sufficient.

There are many plug-ins available contributed by the people who use it.