Currently linkerd expects the cluster to use cluster.local as its base domain name, however more often than not a Kubernetes admin will decide to change that base domain to something else. This shouldn't break deployments, however in the case of linkerd2 it does as every generated configuration, from linkerd install to linkerd inject, assumes connection to happen via cluster.local-based domain names.

A workaround for this issue is to pipe generated configurations into sed 's/cluster\.local/custom\.domain/g. This has been documented shortly in this pull request for the website but I think it would be good for usage sanity to introduce a CLI option like --cluster-domain=… or a similar way to configure the base domain and inject it into the respective template.

Not sure about this but this may also affect linkerd CLI functionality outside of generating templates. (Checks? Proxies?)