It would be useful if check_mysql could check the maximum time a thread
has been in the "Locked" state.  A MySQL bug or a huge query could
effectively stop MySQL being useful, but not prevent the monitor plugin
from connecting and therefor not generate a nagios alert.

Attached is a patch to add a -l/--locktime option to specify the maximum
time thread can be in the 'Locked' state before causing the plugin to
return CRITICAL.