[hotfix] fixup mr !324: mount vhost.d correctly on signalboost prod

* symptoms:
  * prometheus returning no data from signald scrape job
  * ngix logs show: `4132 no live upstreams while connecting to
    upstream, client: REDACTED, server: signald.signalboost.info`
* suspected cause: metrics docker compose vhost.d ip whitelisting
  works because we mount a volume from the filesystem instead of
  using a namded volume, but in signalboost prod configs we still
  use a (vestigal) named volume from example configs
* attemped fix: mount the `vhost.d` directory from the repo instead of using the
  named volume