When you have an application that queries big tables in your database it may happen that some inefficiency querying the database can slow down the whole application a lot. In this cases it is useful to get a log of the queries the database is receiving.
I had this issue while publishing with GeoServer road segments with speed attributes that change over time depending on traffic. The table has a lot of records and the application was struggling. In order to log the queries you just have to set log_statement to ‘all’ in postgresql.conf. But this time the PostgreSQL instance was in a docker container running the Kartoza PostGIS image so you don’t have direct access to the postgresql.conf.
What I did was this:
Copy their postgresql.conf to some folder, with this command:
docker cp <container>:/etc/postgresql/9.3/main/postgresql.conf /var/pg/postgresql.conf
This allowed me to see easily that the web application was querying all the temporal instances of the speed measures at startup, which was killing the database server. Now it is fixed and I hope to post about the application soon.