pgstats: unhandled exception if PostgreSQL cluster is down

[hfriedr]

If the PostgreSQL database cluster is down (or has been deleted), pgstats in PyMunin 0.9.27 responds with an unhandled exception. Is that a design decision, or could pgstats perhaps exit a little more gracefully?

The unhandled Python exception is detected by the abrt daemon, so every five minutes the unhandled Python exception is logged to /var/log/messages:

Jul 3 13:50:38 lvn-itd-571 abrt: detected unhandled Python exception in '/etc/munin/plugins/pgstats-5600'
Jul 3 13:50:38 lvn-itd-571 abrtd: New client connected
Jul 3 13:50:38 lvn-itd-571 abrt-server[33158]: Saved Python crash dump of pid 33149 to /var/spool/abrt/pyhook-2014-07-03-13:50:38-33149
Jul 3 13:50:38 lvn-itd-571 abrtd: Directory 'pyhook-2014-07-03-13:50:38-33149' creation detected
Jul 3 13:50:39 lvn-itd-571 abrt: detected unhandled Python exception in '/etc/munin/plugins/pgstats-5600'
Jul 3 13:50:40 lvn-itd-571 abrtd: New client connected
Jul 3 13:50:40 lvn-itd-571 abrt-server[33165]: Not saving repeating crash in '/etc/munin/plugins/pgstats-5600'

(pgstats-5600 is a symbolic link pointing to /usr/share/munin/plugins/pgstats, of course. 5600 used to be the port number used by that former PostgreSQL instance.)