Hi,
by aboard you mean like visitors come across all different regions ?
in that case, it is better not to set server IP to node IP , but leave it to your origin server IP
here is an example :
Spanish visitor -> QC - Madrid node -> your Milan server
Franch visitor -> QC - Paris node -> your Milan server
Italian visitor -> QC - Milan node -> your Milan server
so this is how the communication works , assume you run crawler on your origin server , then , after crawler has run , the situation will be like this :
Spanish visitor -> QC - Madrid node / cache miss -> your Milan server / cache hit
Franch visitor -> QC - Paris node / cache miss -> your Milan server / cache hit
Italian visitor -> QC - Milan node / cache miss -> your Milan server / cache hit
and this is scenario if you set crwaler to run on Milan node
Spanish visitor -> QC - Madrid node / cache miss -> your Milan server / cache miss
Franch visitor -> QC - Paris node / cache miss -> your Milan server / cache miss
Italian visitor -> QC - Milan node / cache hit
this will speed up the visitors in Italy , but not as good as for other regions
so it is more applicable if most of your visitors are focus on certain specific region , like what I have said in beginning
a workaround is that if your targeted visitors are on certain specific region , you can make crawler to crawl on that node directly instead of origin
if your visitors are not geo-specific , then better to crawler on origin
once origin has cache , it will be served like static files when cache miss on nodes , which is already a lot faster than origin - cache miss case.
