The timeout active statement (assuming I remember correctly, helps to level out the spikes/peaks by reporting every minute vs. 5 or whatever is default. I've also tweaked the cache entries for our environment. to look that up to see what your environment can handle.
I also stopped doing ingress and egress as it was just too much data and [duplicated really...] already have the info in reverse.
ip flow-cache timeout active 1
ip flow-export source Loopback0
ip flow-export version 9
ip flow-export destination a.b.c.d 2055
ip flow-aggregation cache protocol-port
cache entries 2046
cache timeout inactive 200
cache timeout active 45
export destination a.b.c.d 2055
enabled
int vlan 88
ip flow in
int vlan nn
ip flow in
