How are you monitoring the router and switch? Over snmpget or ping etc? Perhaps you would like to show us the cron script. Any errors in the cron log (or exit codes) and in your email (if STDERR has not been redirected to /dev/null)?
MRTG uses snmpget. Does it work on the command line:
# /home/mrtg/mrtg mrtg.cfg
Test whether this works properly. Also, check how long it takes to run. This is because MRTG uses a lock-file mechanism which waits for the previous MRTG session to complete before continuing with a current session.
If it does, your MRTG processes will accumulate because each session will be waiting for the previous session which will be waiting for the previous session ... and so forth to complete.
One way is to uses separate configuration files instead of relying only on one single mrtg.cfg i.e. it can be z1routers.cfg, z2routers.cfg etc. In this case, z1routers.cfg and z2routers.cfg will be run as non-related separate mrtg sessions. You will not face this issue as long as each mrtg session takes less than 10 mins to run.
