You can decide a threshold response time for all queries which you want to investigate .
Lets say you want to investigate all queries which take more than 200ms then set log_min_duration to 200 in postgresql.conf and it will start logging all queries taking more time than ther threshold in the server log.
--Ashish
On 08/19/2010 01:28 PM, Machiel Richards wrote:
Good day everyone.
Firstly I would like to thank everyone in advance for all the help over the last couple of weeks on this mailing list as it has helped me out a lot in my efforts to try and understand Postgresql better.
I am however once again here with some more questions which I hope someone can assist me with as I can not seem to find good documentation on this as yet...
As part of our daily health checks and efforts to assist our client to improve their performance on their postgresql databases, we need to look at the queries that takes the longest to run and then use explain or other methods to try and improve the queries, add indexes where required, etc...
However, I am trying to find out how I will be able to track / see these queries in order for me to be able to investigate these.
the client have a total of 100 databases and to manually monitor these will be too time consuming considering that we have many clients to look after each day.
Can anyone perhaps point me in the right direction here, even if it is in the form of documentation that will assist me in this.
Regards
Machiel
--
With Regards
Ashish Karalkar