Yes, yet agian i have a rather crazy idea,
go with me on this and hopefully it will make a bit
more sense than normal :)
Currently, i have a rather extensive application
that gets a lot of queries, i was wanting to track the
general 'usage' pattern overtime, but of course, that
would require a sort of 'historical view' of pg_stat_activity.
If there was such a thing, then it would let me isolate
hotspots or queries that were often used and allow me
to focus on them first for speed benefits, as well as
allowing me to produce some nifty histograms :)
So, the long and short, is there a pg_stat table
that contains the most 'used' queries and also (i know i
am shooting for the moon here) execution time ?
I know i -could- log_timestamp and log_statement
and then parse out (using perl) the most called ones for
each day and then sort them all and ... but thats a bit of
a pain if there is already such a thing in existence.
pgsql-sql by date
|Next:||From: azwa||Date: 2004-01-14 07:39:34|
|Subject: Re: Insert into:Bad date external representation|
|Previous:||From: Jeff Boes||Date: 2004-01-13 21:07:45|
|Subject: Rule won't let me NOTIFY, no matter how hard I try|