Need help getting multi-process graphs (Metrics Report) working

[Charles Jones]

I'm trying to make multi-host graphs out of the following rrd's:

[hobbit at mon rrd]$ find . -name processes\* |egrep 'web-12|web-13'
./web-12/processes.Apache_app1.rrd
./web-12/processes.Apache_main.rrd
./web-12/processes.Apache_app2.rrd
./web-12/processes.Apache_app4.rrd
./web-12/processes.Apache_app3.rrd
./web-12/processes.Apache_app3.rrd
./web-13/processes.Apache_app1.rrd
./web-13/processes.Apache_main.rrd
./web-13/processes.Apache_app2.rrd
./web-13/processes.Apache_app4.rrd
./web-13/processes.Apache_app3.rrd
./web-13/processes.Apache_app3.rrd

The rrds are simple process counts with "count" as the DS:
[hobbit at mon rrd]$ rrdtool dump ./web-12/processes.Apache_app1.rrd |grep name
                <name> count </name>

Here is my hobbitgraph.cfg definition:

[processes-multi]
        FNPATTERN processes.(.+).rrd
        TITLE Multi-host Process counts
        YAXIS #
        DEF:cnt at RRDIDX@=@RRDFN@:count:AVERAGE
        -l 0 -A
        LINE2:cnt at RRDIDX@#@COLOR@:@RRDPARAM@
        GPRINT:cnt at RRDIDX@:LAST: \: %10.1lf (cur)
        GPRINT:cnt at RRDIDX@:MAX: \: %10.1lf (max)
        GPRINT:cnt at RRDIDX@:MIN: \: %10.1lf (min)
        GPRINT:cnt at RRDIDX@:AVERAGE: \: %10.1lf (avg)\n

This looks to me like it should work, at least assuming that there are 
the same number of .rrd per host (there are), but I'm getting a broken 
graph image when I try to do a metrics report for multi-host processes 
counts.  I have found that if I do a metrics report on a SINGLE host, 
the graph displays properly.  Any idea what could be wrong?