Policy Routing for some web page

hi
i ma connect to internet with 2 ISP’s first one with public IP( isp1) other one with local ip (isp2). i want all traffic that destined for web page (wwww.banesco.com) goes to isp2 .
i defined L7 protocol for this web page{ regexp[^.+(banesco.com).*$ )]} and un mangle with Mark-routing for this L7 protocol.then defines default gateway for this routing marked.

L7 protocol name= los bancos regexp[^.+(banesco.com).*$ )]
chain=prerouting action=mark-routing new-routing-mark=routing Bancos passthrough=no layer7-protocol=Los Bancos src-address=xx.xx.xx.0/xx log=no log-prefix=“”

ip routes :
distance=1 dst 0.0.0.0/0 gateway=186.xx.xx.xx
distance=1 dst 0.0.0.0/0 gateway=192.xx.xx.xx routing-mark=“routing Bancos”

but dose not works ..
trace route to that web page result incorrect .
tracert www.banesco.com goes to 186.xx.xx.xx

i need help .

It doesn’t work like this at all. Packets don’t use hostnames, they use numeric addresses acquired by client by resolving hostname. So it has zero chance to work with traceroute. And even though it might look more hopeful for web browsing, because there you can find hostname, it can’t work either, because it’s not in first packet and after that it’s too late to route connection elsewhere.

If it was simple website without external dependecies, you could add hostname in address list and RouterOS would resolve it for you and keep it updated. Then you’d mark routing based on this (dst-address-list=). Unfortunately, this specific website seems to be downloading dependencies from everywhere else (see developer console in web browser), so you would have to add several other hostnames, but there’s no guarantee that they won’t change to something completely different.

Thank you
you are right .

what is the solutions for this type of policy routing ???

Not everything has solution. It may seem simple, it’s just one website, but it can (and in this case does) consist of resources from different servers and domains.

You can do the thing with address list, but it will only cover initial request to www.banesco.com. Then it downloads tons of scripts from banesco-prod-2019.s3.amazonaws.com and judging by the name, it’s specific to this site, so you’d have to add that too. But you don’t know if this address will be the same tomorrow. And it’s still not the end, there are resources from other sites too. And it’s just index page, there can be many other domains used in other sections. And it can change at any time.

So the first step is to define why exactly you want to do this and go from there. If it’s because some access restrictions, maybe you could find (using trial & error) what part of site does that and use the address list for just that hostname.