Okhan -- for historical info you should consider use of the history database.
Keep in mind that Netezza will restart itself automatically for many reasons to prevent data issues, etc. When this happens the host database process is restarted (and possibly) not the SPUs.
So in the grandeur scheme of things I think your looking for uptime as a form of measurement for management reporting i.e. 99.99% uptime, stability, etc.
I suggest you look at the number of stop/starts and duration of downtime, etc. and then use this to calculate the total uptime & downtime.
I have found that management would like a detail log as to when & why the Netezza server was unavailable. So you may want to account for this too.
Also, I would suggest writing a bash script to collect start/stop information daily along with scheduled maintenance and other reasons for restart.
Here is a monthly uptime calc formula I've used before ...
Minutes = 43200 (30 days * 24 hours * 60 minutes)
A 99.9% uptime would allow for 4.32 mins of downtime i.e. ((1 - .9999) * 43200 )
BTW -- In my experience over the last 4+ years Netezza has improved uptime to almost 100% ... which is way better than most database systems and appliances can claim.
Jeff Mendonca
