Open scareything opened 2 years ago
To clarify, $dst_host
would match output of the `hostname' command of a Device, correct? If so, PERFECT!
How would devices with no domain (p2p hosts) be identified beyond their hostname? A scenario might be that there are two hosts named frontdeskpc
in two different networks. Is there any conflict? They can be further distinguished by subnet or tag or attributes?
It can be easy to forget that ziti "hostnames" are completely fictional as far as the hosts are concerned - that is unless you intentionally align the system hostnames with your ziti hostnames. When using $dst_hostname
and the like, the ziti hostnames are just the addresses that you use to access other devices through the openziti overlay.
So you're free to line up system hostnames with ziti identities, but that alignment is really just a coincidence that you created (and probably need to maintain and keep straight). If you do this, and the hostnames are not sufficient to uniquely identify each device, then you'd probably quickly find that the openziti controller does not allow identities to have the same name.
Adding some variables that are resolved when evaluating intercept.v1
dialOptions.identity
would make it easier to support multiple remote networks. This follows from discussion on discourse.Proposed variables
$dst_fqdn
$dst_hostname
, but$dst_fqdn
seems more clear$dst_host
$dst_domain
$dst_subdomain
$dst_fqdn
$dst_subdomain[n]
$dst_subdomain
$dst_subdomain[0]
== "dc1",$dst_subdomain[1]
== "region1"$dst_matched_wildcard
$dst_fqdn
matches an intercepted wildcard domain$dst_matched_wildcard
== "dc1.openziti.io"