Design Pattern to keep a connection opened?

Can anyone point me in the right direction for keeping a connection to 
my database opened?

The script I need to write will run 24/7 executing a postgresql stored 
procedure which inserts records into the database.  The problem I think 
I might have is using a standard procedural syntax of:

Open connection
Loop
     Call stored procedure/function

...if my connection gets interrupted somewhere during the loop any 
subsequent iterations will obviously fail, so I need a way of 
reconnecting to the database should the need arise.  Basically something 
like this (I'm just not sure how to actually code this):

my $db_object

sub get_database_object {
     if $db_object exists is functional, return that
     otherwise re-open $db_object and return it
}

sub insert_record {
     my (param1, param2, param3, param4) = @_;
      my $local_db_object = get_database_object();

     # ... code here that utilizes $local_db_object, which should always
     # be functional due to checks performed in get_database_object()
}

Loop
     insert_record("value 1", "value 2", "value 3", "value 4");
End Loop


The main question is, how can I efficiently determine if a db object is 
usable?

Thanks,
Brandon