Not planning to work on it directly, but I am planning to do some larger executor work as one of the next bigger projects, and having ideas about what kind of information people would like to see could make it easier to implement them later.
Not planning to work on it directly, but I am planning to do some larger executor work as one of the next bigger projects, and having ideas about what kind of information people would like to see could make it easier to implement them later.
Just always collecting that information just in case it may get accessed thus isn't really feasible. It'd be good to make it possible to query cheaper information on-demand though (e.g. asking for the EXPLAIN of a query running in another session, without analyze, should be doable with some effort).
You can already set up things in a way that allows to correlate connections / queries with distributed tracing. But it's a more work than it should be. Postgres' pg_stat_activity shows queries, and it can include information that allows to correlate in the connection's 'application_name'.