Fix orchestrator and template manager health checks

[djeebus]

• use provider=nomad explicitly
• grpc is only supported by consul-integrated nomad clusters, use http instead
• add tcp health check for orchestrator proxy port

Before:

After:

---

Note

Set provider="nomad", switch gRPC health checks to HTTP (/health), and add a TCP health check for the orchestrator proxy.

• Nomad jobs:
  • iac/provider-gcp/nomad/jobs/orchestrator.hcl:
    • Service orchestrator: add provider = "nomad"; change health check from grpc to http with path = "/health".
    • Service orchestrator-proxy: add provider = "nomad"; add tcp health check.
  • iac/provider-gcp/nomad/jobs/template-manager.hcl:
    • Service template-manager: add provider = "nomad"; change health check from grpc to http with path = "/health".

^{Written by Cursor Bugbot for commit d6c8455. This will update automatically on new commits. Configure here.}

[djeebus]

This enables the health checks to work. Without it, we default to consul, which doesn't function.

[djeebus]

Don't need this, as we're using HTTP (not GRPC) and it inherits the port of the service block that surrounds the check.

[ValentaTomas]

Just for clarity—it inherits the check from orchestrator-proxy, not the orchestrator grpc, right?

[djeebus]

It inherits the port, that's all:

job "orchestrator" {
    group "client" {
        service {
            name = "orchestrator"
            port = "1234"

            check {
                # inherits `port = "1234"`
            }
        }

        service {
            name = "proxy"
            port = "5678"

            check {
                # inherits `port = "5678"`
            }
        }
    }
}

[ValentaTomas]

Oh, so there are still two healthchecks, right?

[ValentaTomas]

Lgtm, just wanted to wait if @jakubno does not have any notes about some possible bad interaction with the current orchestrator job deploy?