You would set two static routes. First route will point to your first ISP and will be exactly what you use today. The second router will have a higher administrative distance, set it for say 150, and traffic will only follow the second route when the first one goes down.
Primary Link distance = 1
Backup Link distance = 2
Let's say what will happen if the local primary ISP is reachable till there NOC, but having trouble from upstream provider ? the check gateway-ping will only ping then next hop that is either local Router (in case of dsl router) or ISP Gateway (in case of static ip or bridge mode) which will eventually always reachable , so the route will always be considered as UP. (even if the internet is actually Down)
It is recommended to configure an script
that should run every mint or so via scheduler , that should check at least 2 internet hosts (that must always route via the primary link / you can do this in route section), if both hosts are unreachable then the primary link distance should be changed to 3 , so that all request should route via backup link & once the both or one internet hosts are reachable then the script should change the primary link distance back to 1
If you have single masquerade rule (without pointing the OUT interface) then you dont need any further additional rules.