hhvm.conf

# This file should be in /etc/monit/conf.d
check process hhvm with pidfile /var/run/hhvm/pid
  group hhvm
  # Start program calls our custom script from above
  start program = "/usr/local/sbin/start_hhvm.sh" 
  stop program  = "/usr/sbin/service hhvm stop"
  if failed unixsocket /var/run/hhvm/hhvm.sock then restart
  if mem > 400.0 MB for 1 cycles then restart
  if 5 restarts with 5 cycles then timeout

start_hhvm.sh

#!/bin/bash

# This file should be saved in /usr/local/sbin
# It must exist because sometime a crashed process 
# leaves a socket behind, rendering it unstartable
# So what we do is test whether the socket file
# leftover is there and delete it. Also, make sure
# this file is executable: $ chmod +x start_hhvm.sh

SOCKET=/var/run/hhvm/hhvm.sock
if [ -e $SOCKET ];
then
  rm -f $SOCKET
fi

/usr/sbin/service hhvm restart
exit 0;

in start_hhvm.sh

service hhvm start

had to change to

service hhvm restart