I am not a PostgresSQL user but would like to hear from users who have configured their database to use the UTF-8 charset, especially any users using the case and accent insensitive collation.
I am using a open source SQL server at this time using ISO8859_1 and performance is excellent but I recreated the structure using UTF-8 and imported the data in a test environment and the performance was totally unacceptable. Example:
An unoptimized (no index) query under ISO8859_1 runs in about 15ms and took almost 6.5 seconds under UTF-8.
Creating an index did help but performance was still unacceptable. Using the same query under ISO8859_1 now took a reported 0ms while taking 109ms under UTF-8.