Add pyth quorom server #2759
Add pyth quorom server #2759
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
There was a problem hiding this comment.
Pull Request Overview
This PR implements a new Pyth server (named "quorom") that includes a websocket endpoint, REST API endpoints to handle observations, and integration with Pythnet for fetching guardian set information.
- Implements a websocket actor (Subscriber) to manage long-lived connections and deliver update events.
- Adds API endpoints for health checks and handling observation submissions that result in broadcasting VAAs.
- Introduces a Pythnet client to fetch guardian set data and integrates server shutdown via a global exit flag.
Reviewed Changes
Copilot reviewed 9 out of 9 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
| apps/quorom/src/ws.rs | Implements websocket handling and subscriber actor for update events. |
| apps/quorom/src/server.rs | Sets up server state, command-line options, and graceful shutdown. |
| apps/quorom/src/pythnet.rs | Implements Pythnet client functionality to fetch guardian set info. |
| apps/quorom/src/main.rs | Configures tracing and starts the server. |
| apps/quorom/src/api.rs | Provides API endpoints to handle observations and broadcast VAAs. |
| apps/quorom/rust-toolchain | Specifies the Rust toolchain version. |
| apps/quorom/Cargo.toml | Declares the project metadata and dependencies. |
| apps/quorom/.gitignore | Configures ignore rules for Rust build artifacts. |
apps/quorom/src/api.rs
Outdated
| if let Err(e) = state.ws.broadcast_sender.send(UpdateEvent::NewVaa(vaa)) { | ||
| tracing::error!(error = ?e, "Failed to broadcast new VAA"); | ||
| } | ||
| verification_writer.remove(¶ms.body); |
There was a problem hiding this comment.
something that can happen here is that we wait for all these to finish when we shut down. i am somewhat surprised that axum graceful shutdown is not doing it.
There was a problem hiding this comment.
I used task tracker to make sure that we handle all observation before shutting down. Do you think it's not necessary to handle all observations before shutting down?
There was a problem hiding this comment.
This means we are not going to accept new requests, but we will wait until we handle all prv requests that we received
There was a problem hiding this comment.
When we receive a Ctrl-c signal, the WS connection is being dropped immediately, so I just removed the task tracker for simplicity.
I also tested graceful shutdown and it's working fine.
This PR aims to add pyth quorom server