Cacti 0.8.7i polling with cmd.php every minute with 10 processes

When looking at Cacti logfile it appears that polling is happening twice in a 60 second window. One completes in about 5 seconds and one in about 25 seconds:

02/20/2012 04:08:09 PM - SYSTEM THOLD STATS: Time:2.0177 Tholds:278 DownHosts:0
02/20/2012 04:08:07 PM - SYSTEM STATS: Time:5.3997 Method:cmd.php Processes:10 Threads:N/A Hosts:85 HostsPerProcess:9 DataSources:2853 RRDsProcessed:882
02/20/2012 04:07:28 PM - SYSTEM THOLD STATS: Time:1.7841 Tholds:278 DownHosts:0
02/20/2012 04:07:26 PM - SYSTEM STATS: Time:24.7145 Method:cmd.php Processes:10 Threads:N/A Hosts:85 HostsPerProcess:9 DataSources:2853 RRDsProcessed:882

I can't imagine that this is right. Can anyone suggest where to begin troubleshooting this?

Update:

I've installed Spine and have the same issue. According to my scheduled cron jobs in Webmin I have one command running every 60 seconds:

/usr/bin/php /var/www/html/cacti/poller.php > /dev/null 2>&1

I think it's not so big problem..
Your poller started at every minute, and finished within 60 seconds.

Where I was concerned was that it ran two pollers in that one minute window. I changed the Maximum Concurrent Poller Processes to '1' and that resolved the problem. I also adjusted the spine specific parameters to increase the Maximum Threads per Process to 50, Number of PHP Script Servers to 10, Script and Script Server Timeout Value to 40, and The Maximum SNMP OID's Per SNMP Get Request to 100. I've not seen a "best practices" guide but the poller completes in about 24 seconds. Is thaht normal even for using spine?

SYSTEM STATS: Time:23.2544 Method:spine Processes:1 Threads:50 Hosts:84 HostsPerProcess:84 DataSources:2766 RRDsProcessed:867