Per-domain processing?

[Alex]

Hi,
I'm using amavis with postfix-3.10 and spamassassin and would like to
be able to apply different policies for different domains, but I'm
unsure how to do that.

I'm using postfix in a multi-instance configuration, with each
instance processing mail for a different domain (or set of domains)
and would like to be able to have one amavisd process still handle all
postfix instances. I suppose I could otherwise use transport maps with
postfix and different amavisd processes, each of which with a
different port and different config.

Currently I have @local_domains_maps set to all domains for which it
should process mail, but I'd like to be able to control virus/spam
scanning, blocklist/allowlist, and other features on a per-domain
basis. Is this possible?

I'm aware of $policy_bank but isn't that based on IP address? Can I
set @client_ipaddr_policy based on domain rather than IP? Or perhaps
there's some sort of @client_domain_policy?

Also, Example 3 in the policy banks section from the main doc
(https://opensource.apple.com/source/amavisd/amavisd-114/amavisd/amavisd-new-2.4.4/README_FILES/amavisd-new-docs.html#pbanks-ex)
seems to be closest to what I need, but I'm still not understanding.
Perhaps it's possible to have one amavisd process listen on multiple
sockets with different config files include files, each of which has a
different configuration?

Ideas greatly appreciated.