openai/codex
1
0
Fork 0
mirror of https://github.com/openai/codex.git synced 2026-10-20 17:38:45 +00:00
Code Issues Packages Projects Releases Wiki Activity

Trace harness-side Noise handshake failures (#55406)

Browse Source 
## Why

The handshake readiness channel only signals success, so a dropped sender does not explain why the handshake failed.

## What changed

Emit structured `codex.exec_server.noise_handshake_failed` warnings before harness handshake failures drop the readiness sender, including a failure reason and environment, executor registration, and stream identifiers. Emit the same event on handshake timeout with `timeout_ms`.

Preserve the current tracing span when the `noise_relay.stream` debug span is disabled so the relay task retains its trace context.

GitOrigin-RevId: 0b1314d5237b7c77a854c2c0a966391bf856e75d
This commit is contained in:
alishobeiri-oai
2026-10-19 23:56:17 +00:00
committed by copyberry
parent 79a19cbb30
commit c937a6eb7e
2 changed files with 45 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
codex-rs/exec-server/src/client_transport.rs
+16 -3
View File
@@ -725,9 +725,22 @@ impl ExecServerClient {
Ok(Err(_)) => Err(ExecServerError::Disconnected(
"Noise harness handshake failed before connection became ready".to_string(),
)),
Err(_) => Err(ExecServerError::InitializeTimedOut {
timeout: initialize_timeout,
}),
Err(_) => {
tracing::event!(
name: "codex.exec_server.noise_handshake_failed",
target: "codex_otel.trace_safe",
tracing::Level::WARN,
event.name = "codex.exec_server.noise_handshake_failed",
noise_event = "handshake",
noise_outcome = "failed",
noise_side = "harness",
noise_reason = "timeout",
timeout_ms = initialize_timeout.as_millis() as u64,
);
Err(ExecServerError::InitializeTimedOut {
timeout: initialize_timeout,
})
}
}
}
codex-rs/exec-server/src/noise_relay/harness.rs
+29 -1
View File
@@ -101,7 +101,8 @@ where
let (incoming_tx, incoming_rx) = mpsc::channel(CHANNEL_CAPACITY);
let (disconnected_tx, disconnected_rx) = watch::channel(false);
let (handshake_ready_tx, handshake_ready_rx) = oneshot::channel();
let stream_span = tracing::debug_span!("noise_relay.stream", noise_side = "harness",);
let stream_span =
tracing::debug_span!("noise_relay.stream", noise_side = "harness",).or_current();
debug!(
environment_id,
executor_registration_id, stream_id, "Noise harness relay details"
@@ -110,6 +111,22 @@ where
let websocket_task = tokio::spawn(async move {
let mut handshake_ready_tx = Some(handshake_ready_tx);
let mut websocket = stream;
// Readiness only carries success; record failures before that signal is dropped.
let record_handshake_failure = |reason: &'static str| {
tracing::event!(
name: "codex.exec_server.noise_handshake_failed",
target: "codex_otel.trace_safe",
tracing::Level::WARN,
event.name = "codex.exec_server.noise_handshake_failed",
noise_event = "handshake",
noise_outcome = "failed",
noise_side = "harness",
noise_reason = reason,
environment_id,
executor_registration_id,
stream_id,
);
};
// Bind the Noise transcript to the exact environment registration and
// virtual relay stream before emitting any handshake bytes. A captured
@@ -124,6 +141,7 @@ where
) {
Ok(handshake) => handshake,
Err(error) => {
record_handshake_failure("handshake_start");
send_disconnected(
&incoming_tx,
&disconnected_tx,
@@ -149,6 +167,7 @@ where
.await
.is_err()
{
record_handshake_failure("websocket_write");
let _ = disconnected_tx.send(true);
return;
}
@@ -158,6 +177,7 @@ where
// create a plaintext or unauthenticated application path.
let mut transport = loop {
let Some(incoming_message) = websocket.next().await else {
record_handshake_failure("websocket_ended");
send_disconnected(
&incoming_tx,
&disconnected_tx,
@@ -168,6 +188,7 @@ where
let message = match incoming_message {
Ok(Message::Binary(payload)) => payload,
Ok(Message::Close(_)) => {
record_handshake_failure("websocket_close");
send_disconnected(
&incoming_tx,
&disconnected_tx,
@@ -177,6 +198,7 @@ where
}
Ok(Message::Ping(_) | Message::Pong(_) | Message::Frame(_)) => continue,
Ok(Message::Text(_)) => {
record_handshake_failure("unexpected_text");
send_disconnected(
&incoming_tx,
&disconnected_tx,
@@ -185,6 +207,7 @@ where
return;
}
Err(error) => {
record_handshake_failure("websocket_read");
send_disconnected(
&incoming_tx,
&disconnected_tx,
@@ -198,6 +221,7 @@ where
let frame = match decode_relay_message_frame(message.as_ref()) {
Ok(frame) => frame,
Err(error) => {
record_handshake_failure("invalid_frame");
send_disconnected(
&incoming_tx,
&disconnected_tx,
@@ -215,6 +239,7 @@ where
let response = match frame.into_handshake_payload() {
Ok(response) => response,
Err(error) => {
record_handshake_failure("invalid_response");
send_disconnected(
&incoming_tx,
&disconnected_tx,
@@ -236,6 +261,7 @@ where
break transport;
}
Err(error) => {
record_handshake_failure("handshake_finish");
send_disconnected(
&incoming_tx,
&disconnected_tx,
@@ -246,6 +272,7 @@ where
}
}
Ok(RelayFrameBodyKind::Reset) => {
record_handshake_failure("relay_reset");
send_disconnected(
&incoming_tx,
&disconnected_tx,
@@ -259,6 +286,7 @@ where
| RelayFrameBodyKind::Heartbeat,
) => {}
Ok(RelayFrameBodyKind::Data) | Err(_) => {
record_handshake_failure("unexpected_frame");
send_disconnected(
&incoming_tx,
&disconnected_tx,