Differentiate vm_type for tagged worker and add new ops file for setting proxy for tagged worker

[brightzheng100]

Two changes in this PR:

1. Use a dedicated variable tagged_worker_vm_type for setting the tagged worker

It's common to have different vm types for no-tag worker pool and tagged worker.

For example, we may use xlarge vm type for default shareable workers, while setting a xlarge_disks one for tagged worker for BBR purposes.

2. Create a new ops file for setting the http(s) proxy for tagged worker

The current http-proxy.yml can't support proxy setting for tagged worker. So we need a new one for it.

As a result, if we're going to spin up a cluster with shareable worker pool and tagged worker, and (optionally) to set the http(s) proxy for it, we can do something like:

bosh deploy -d concourse cluster/concourse.yml \
  -l versions.yml \
  ...
  -o cluster/operations/http-proxy.yml \
  ...
  -o cluster/operations/tagged-worker.yml \
  -o cluster/operations/http-proxy-tagged-worker.yml \
  ...
  -v http_proxy_url=https://myproxy.com \
  -v https_proxy_url=https://myproxy.com \
  -v no_proxy='["localhost", "127.0.0.1", "example.com", "domain.com:8080"]'
  ...
  -v worker_tag=bbr \
  -v tagged_worker_vm_type=xlarge_disks \
  ...

Note: the newly added http-proxy-tagged-worker.yml fully respects the variables defined in http-proxy.yml and no new variables are added.

[deniseyu]

Seems reasonable. Thanks for the PR @brightzheng100!