Support health-check aliases

Inter alia, our Nomad clusters run a two-tier application suite where the health of a front-end service indicates the healthiness of a back-end application. The back-end application does not expose any services. It would be convenient if we could instruct Nomad that the healthiness of the back-end allocation is indicated by some service exposed by another allocation, such that `check_restart` would trigger a restart of the back-end container as well.

Currently, we work around this issue by deploying both applications as part of the same task group, so that the back-end is eventually restarted after front-end service failure once the re-scheduling policy triggers.

I’m not yet familiar with Consul alias-type health-checks, but perhaps this feature could be achieved integrating these into Nomad.

The issue names no files or tests. Start by reviewing how Nomad handles check_restart and how Consul alias-type health checks work, then trace how health checks relate to allocations and task groups. Done means an external service health check can restart the dependent back-end allocation without placing both applications in one task group.

Written by the indexing model from the issue text.