On Mar 22, 2006, at 11:11 AM, Scott Marlowe wrote:

> Would welcome any suggestions for a better query to check the  
> health of
> the set.

My simple test, which I use from a nagios monitoring plugin, is this:

SELECT date_part('epoch',st_lag_time)::INTEGER FROM _$ 
{cluster}.sl_status WHERE st_origin=$origin_node AND st_received= 
$subscriber_node

It gives me the number of seconds behind the subscriber is, which is  
for me the most important measure.  The lag_time is based on last  
event timestamp.  I suspect both queries give basically the same info  
in the end.