I like how postgres backends modify the 'command' display in a (unix)
process listing to show what database they are connected to, and the
query being executed.
Has any thought been put into being able to make this configurable per
backend connection (maybe via a SET variable)? I'm thinking it would
be cool if a user process could provide its own status there,
identifying application-specific information. Similar to how Java
let's you set the "name" of a Thread (which sadly doesn't appear in a
process listing, but oh well).
I realize this could open up some security concerns by allowing us to
expose more information than we should, but for those of us that want
to shoot ourselves in the foot, it might be a useful feature.
Any thoughts?
eric