As a first step, before relaunching the query with a new plan I would be
happy to be able to get information about sql queries having wrong
estimates.
Maybe thoses SQL queries could be collected using something similar to
"auto_explain" module (tracing finished or cancelled queries).
If the "corrected plan" taking into account real cardinalities was proposed
(with some advices on how to get it) it would be a great tuning adviser ;o)
--
Sent from: http://www.postgresql-archive.org/PostgreSQL-performance-f2050081.html