Add gRPC server inference SSH example

[cretz]

What changed

• New example follow-on from Add inference SSH tune-vllm-args example #581.
• Full README.md walking through how to use it (manually tested).
• Adds inference-ssh/grpc-server, an inference-SSH example that deploys a Python gRPC server and shows editing its code on a running replica and restarting it in place, no redeploy.
• Server is a Greeter service plus the standard gRPC health check service (how Baseten judges gRPC health), packaged as grpc_greeter in a self-contained uv project under project/.
• config.yaml runs it as an opaque docker_server over runtime.transport.kind: grpc on port 50051, CPU-only, with remote_ssh enabled: stock python:3.12-slim, external_package_dirs ships the project into /packages, uv sync --frozen --no-dev installs locked runtime deps editable, and entr restarts the server on any package-file change.
• Includes committed protobuf stubs plus uv run grpc-greeter-codegen to regenerate them, and a uv run grpc-greeter-client that calls the deployed model over gRPC.
• Links the new example from inference-ssh/README.md.

[Copilot]

Pull request overview

Adds a new inference-ssh/grpc-server example demonstrating an in-container edit/restart loop over SSH for a Python gRPC server on Baseten, including a walkthrough and a self-contained uv project with generated protobuf stubs.

Changes:

• Add a new grpc-server inference-SSH example with full README walkthrough, Truss config.yaml, and a simple Greeter + gRPC health-check server.
• Add a uv-managed Python project (grpc_greeter) including client, codegen script, and checked-in generated stubs.
• Link the new example from the inference-ssh index README.

Reviewed changes

Copilot reviewed 10 out of 12 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
inference-ssh/README.md	Adds link to the new grpc-server SSH dev-loop example.
inference-ssh/grpc-server/README.md	Step-by-step deployment, calling, SSH editing, and restart walkthrough.
inference-ssh/grpc-server/config.yaml	Truss config to run an opaque gRPC docker_server with remote_ssh and entr restarts.
inference-ssh/grpc-server/project/pyproject.toml	Defines the uv project, runtime deps, and uv run scripts for client/codegen.
inference-ssh/grpc-server/project/uv.lock	Locks Python dependencies for the example project.
inference-ssh/grpc-server/project/greeter.proto	Proto contract for the Greeter service.
inference-ssh/grpc-server/project/src/grpc_greeter/init.py	Package marker/docstring for grpc_greeter.
inference-ssh/grpc-server/project/src/grpc_greeter/server.py	gRPC server implementation including gRPC health-check registration.
inference-ssh/grpc-server/project/src/grpc_greeter/client.py	Client for calling the deployed model over gRPC with Baseten metadata.
inference-ssh/grpc-server/project/src/grpc_greeter/codegen.py	Script to regenerate protobuf/gRPC Python stubs via grpc_tools.protoc.
inference-ssh/grpc-server/project/src/grpc_greeter/greeter_pb2.py	Checked-in generated protobuf Python stub.
inference-ssh/grpc-server/project/src/grpc_greeter/greeter_pb2_grpc.py	Checked-in generated gRPC Python stub.

Files not reviewed (1)

• inference-ssh/grpc-server/project/src/grpc_greeter/greeter_pb2.py: Generated file

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[marius-baseten]

Nit: add request type, I assume this is HelloRequest?.

[cretz]

Annotations added

[marius-baseten]

We have already another SSH example. Is the point of this one to show how to do gRPC + SSH in combination specifically? Is SSH with gRPC different than with the vllm example? I'm a bit confused with of the two topics SSH and gRPC is the focus point of this one here .

[cretz]

Is SSH with gRPC different than with the vllm example?

No, not much different and I agree it may not deserve its own sample, but this was an explicit request by @squidarth to be a separate sample regardless.