Support for Supabase/Supavisor

Extracted from @andreirtaylor in #205:

I'm assuming that this is related, but I have not been able to get river to work with supabase they use a proprietary connection pooling system called supavisor.

The error output from river is below when using session pooling.

Happy to help with any testing here and or opening up a new issue.

time=2024-02-25T11:52:38.976-05:00 level=ERROR msg="Scheduler: Error scheduling jobs" error="error deleting completed jobs: ERROR: prepared statement \"stmtcache_9233d4c846c6a2a54af178e1e500878088ef7296ba8c6bf0\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:39.275-05:00 level=ERROR msg="error attempting reelection" elector.err="ERROR: prepared statement \"stmtcache_e924bba8ef07b260df276cc58553fdf77a5d2ac1321679b8\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:39.504-05:00 level=ERROR msg="producer: Error fetching jobs" err="ERROR: prepared statement \"stmtcache_533cb1679bed1f4917f01ee331deb9347cca82532ad1cad9\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:40.416-05:00 level=ERROR msg="producer: Error fetching jobs" err="ERROR: prepared statement \"stmtcache_533cb1679bed1f4917f01ee331deb9347cca82532ad1cad9\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:41.393-05:00 level=ERROR msg="producer: Error fetching jobs" err="ERROR: prepared statement \"stmtcache_533cb1679bed1f4917f01ee331deb9347cca82532ad1cad9\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:42.405-05:00 level=ERROR msg="producer: Error fetching jobs" err="ERROR: prepared statement \"stmtcache_533cb1679bed1f4917f01ee331deb9347cca82532ad1cad9\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:43.381-05:00 level=ERROR msg="producer: Error fetching jobs" err="ERROR: prepared statement \"stmtcache_533cb1679bed1f4917f01ee331deb9347cca82532ad1cad9\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:43.977-05:00 level=ERROR msg="Scheduler: Error scheduling jobs" error="error deleting completed jobs: ERROR: prepared statement \"stmtcache_9233d4c846c6a2a54af178e1e500878088ef7296ba8c6bf0\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:44.394-05:00 level=ERROR msg="producer: Error fetching jobs" err="ERROR: prepared statement \"stmtcache_533cb1679bed1f4917f01ee331deb9347cca82532ad1cad9\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:44.644-05:00 level=ERROR msg="error attempting reelection" elector.err="ERROR: prepared statement \"stmtcache_e924bba8ef07b260df276cc58553fdf77a5d2ac1321679b8\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:45.374-05:00 level=ERROR msg="producer: Error fetching jobs" err="ERROR: prepared statement \"stmtcache_533cb1679bed1f4917f01ee331deb9347cca82532ad1cad9\" already exists (SQLSTATE 42P05)"
time=2024-02-25T11:52:46.418-05:00 level=ERROR msg="producer: Error fetching jobs" err="ERROR: prepared statement \"stmtcache_533cb1679bed1f4917f01ee331deb9347cca82532ad1cad9\" already exists (SQLSTATE 42P05)"

Start by reproducing the reported Supavisor session-pooling failure using the prepared-statement errors in this issue. No source file or test is identified in the report, so trace the PostgreSQL connection setup first and determine the compatibility change and regression test needed. Done means River can operate through Supavisor session pooling without these errors.