On Wed, Dec 24, 2008 at 9:31 AM, Kevin Kempter
<kevink@consistentstate.com> wrote:
> Hi All;
>
> I'm working on a vacuum monitor script to alert us when tables start to grow
> out of control per dead space. Initially I excluded the system catalog
> schemas (pg_catalog, pg_toast, etc) however I wonder if maybe I should
> monitor these as well. PLus I'm looking for any input as to specifically
> what to look for - I've calculated a 'fill factor' which shows the average
> page fill as a percentage but I wonder what other metrics may help monitor
> possible problem tables effeciently..
It's a good idea to keep an eye on system catalogs, especially if you
have a fair bit of DDL going on in your usage. I'd take a look at
what the check_postgresql.pl script does. No need to reproduce all
that work on your own.