performance RB1100AHx2

Hello, I currently have a symmetric connection of 270 MG and a rb1100ahx2.
This team makes NAT + dhcp + queue simple (350 Queues)
It has a bridge with ports 1 to 10.
It has 10 firewall filter rules and 3 mangle rules
I have 2 wan configured and their routing rules.
The CPU consumption reaches a maximum of 54%, but I notice that in peak hours everything gets slower. Will the team be too small?

en español:

Hola, actualmente tengo una conexion simetrica de 270 MG y un rb1100ahx2.
Este equipo hace NAT + dhcp + queue simple (350 Colas)
Tiene un bridge con los puertos 1 al 10.
Tiene 10 reglas de filter en firewall y 3 reglas de mangle.
Tengo 2 wan configuradas y sus reglas de ruteo.
El consumo de CPU llega maximo a 54 %, pero noto que en horas picos se pone todo mas lento. Sera que queda chico el equipo?

I can only guess not properly configured queues…
Do you have a global parent in your queues?
Have you set correct limit-at and priority attributes ?

it is an ISP and the queues are automatically agreed from the dhcp allocation, with the parameter rate.
rate-limit = 1M / 3M
Queues are created automatically every time someone connects to the Dhcp

and they look like this
target = 172.16.10.67 / 32 parent = none packet-marks = “” priority = 8/8 queue = default-small / default-small limit-at = 1M / 3M max-limit = 1M / 3M
burst-limit = 0/0 burst-threshold = 0/0 burst-time = 0s / 0s bucket-size = 0.1 / 0.1

thanks

The information are not enough so that i can help… It is really difficult to find what the problem might be… It can be anything…