feat(tunnel): 添加流帧分发超时机制和结构化请求日志

- dispatcher: 为 stream handler 的帧接收添加超时保护, 防止单个
  handler 阻塞 WebSocket 读循环; 超时后发送 StreamError 并清理流
- stream_handler: 在所有请求完成和错误路径添加结构化日志, 记录
  method/host/path/status/duration 等关键信息
- 将常规隧道连接/断开日志从 info 降级为 debug, 减少日志噪音

apps/aether-proxy/src/tunnel/client.rs

@@ -8,7 +8,7 @@ use tokio::sync::watch;
 use tokio_tungstenite::tungstenite::client::IntoClientRequest;
 use tokio_tungstenite::tungstenite::http;
 use tokio_tungstenite::tungstenite::protocol::WebSocketConfig;
-use tracing::{debug, info, warn};
+use tracing::{debug, warn};
 
 use crate::state::{AppState, ServerContext};
 
@@ -34,7 +34,7 @@ pub async fn connect_and_run(
     drain: watch::Receiver<bool>,
 ) -> Result<TunnelOutcome, anyhow::Error> {
     let ws_url = build_tunnel_url(server);
-    info!(url = %ws_url, conn = conn_idx, "connecting tunnel");
+    debug!(url = %ws_url, conn = conn_idx, "connecting tunnel");
 
     // Build WebSocket request with auth headers
     let mut request = ws_url.clone().into_client_request()?;
@@ -124,7 +124,7 @@ pub async fn connect_and_run(
         .config
         .tunnel_ping_interval()
         .expect("validated config should resolve tunnel ping interval");
-    info!(
+    debug!(
         conn = conn_idx,
         tcp_keepalive_secs = state.config.tunnel_tcp_keepalive_secs,
         tcp_nodelay = state.config.tunnel_tcp_nodelay,
@@ -213,7 +213,7 @@ pub async fn connect_and_run(
         let _ = tokio::time::timeout(Duration::from_secs(35), writer_handle).await;
     }
 
-    info!("tunnel disconnected");
+    debug!("tunnel disconnected");
     Ok(outcome)
 }
 

apps/aether-proxy/src/tunnel/dispatcher.rs

@@ -19,6 +19,13 @@ use super::protocol::{decompress_if_gzip, Frame, MsgType, RequestMeta};
 use super::stream_handler;
 use super::writer::FrameSender;
 
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+enum StreamDispatchStatus {
+    Delivered,
+    Closed,
+    TimedOut,
+}
+
 /// Run the dispatcher loop, reading from the WebSocket stream.
 pub async fn run<S>(
     state: Arc<AppState>,
@@ -97,7 +104,7 @@ where
             Message::Ping(_) => continue,
             Message::Pong(_) => continue,
             Message::Close(_) => {
-                info!("received WebSocket close");
+                debug!("received WebSocket close");
                 break None;
             }
             _ => continue,
@@ -209,12 +216,19 @@ where
             }
 
             MsgType::RequestBody => {
-                if let Some(tx) = streams.get(&frame.stream_id) {
+                if let Some(tx) = streams.get(&frame.stream_id).cloned() {
                     let is_end = frame.is_end_stream();
                     let sid = frame.stream_id;
-                    let _ = tx.send(frame).await;
-                    if is_end {
+                    let dispatch = dispatch_stream_frame(&tx, frame).await;
+                    if is_end || dispatch != StreamDispatchStatus::Delivered {
                         streams.remove(&sid);
+                        if dispatch == StreamDispatchStatus::TimedOut {
+                            try_send_stream_error(
+                                &frame_tx,
+                                sid,
+                                "proxy request body dispatch stalled",
+                            );
+                        }
                         if draining && streams.is_empty() {
                             info!("tunnel drained after request body completion");
                             break None;
@@ -226,7 +240,7 @@ where
             MsgType::StreamEnd | MsgType::StreamError => {
                 // Client-side cancellation or end
                 if let Some(tx) = streams.remove(&frame.stream_id) {
-                    let _ = tx.send(frame).await;
+                    let _ = dispatch_stream_frame(&tx, frame).await;
                     if draining && streams.is_empty() {
                         info!("tunnel drained after stream termination");
                         break None;
@@ -280,6 +294,59 @@ where
     }
 }
 
+async fn dispatch_stream_frame(tx: &mpsc::Sender<Frame>, frame: Frame) -> StreamDispatchStatus {
+    let stream_id = frame.stream_id;
+    match tokio::time::timeout(stream_frame_dispatch_timeout(), tx.send(frame)).await {
+        Ok(Ok(())) => StreamDispatchStatus::Delivered,
+        Ok(Err(_)) => {
+            warn!(
+                stream_id,
+                "stream handler channel closed while dispatching tunnel frame"
+            );
+            StreamDispatchStatus::Closed
+        }
+        Err(_) => {
+            warn!(
+                stream_id,
+                timeout_ms = stream_frame_dispatch_timeout().as_millis(),
+                "stream handler channel blocked while dispatching tunnel frame"
+            );
+            StreamDispatchStatus::TimedOut
+        }
+    }
+}
+
+/// Bound how long a single stream handler is allowed to block the shared
+/// WebSocket read loop while receiving request-body frames.
+fn stream_frame_dispatch_timeout() -> Duration {
+    #[cfg(test)]
+    {
+        Duration::from_millis(25)
+    }
+
+    #[cfg(not(test))]
+    {
+        Duration::from_secs(5)
+    }
+}
+
+fn try_send_stream_error(frame_tx: &FrameSender, stream_id: u32, message: &'static str) {
+    if frame_tx
+        .try_send(Frame::new(
+            stream_id,
+            MsgType::StreamError,
+            0,
+            Bytes::from(message),
+        ))
+        .is_err()
+    {
+        warn!(
+            stream_id,
+            "writer channel full, StreamError dropped while aborting stalled stream"
+        );
+    }
+}
+
 /// Wait for all active stream handlers to finish (with a timeout).
 async fn drain_handlers(handles: Vec<JoinHandle<()>>) {
     if handles.is_empty() {
@@ -294,3 +361,61 @@ async fn drain_handlers(handles: Vec<JoinHandle<()>>) {
     })
     .await;
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use aether_runtime::bounded_queue;
+
+    #[tokio::test]
+    async fn dispatch_stream_frame_times_out_when_handler_stops_draining() {
+        let (tx, mut rx) = mpsc::channel::<Frame>(1);
+        tx.send(Frame::new(
+            7,
+            MsgType::RequestBody,
+            0,
+            Bytes::from_static(b"first"),
+        ))
+        .await
+        .expect("first frame should enqueue");
+
+        let stalled_send = tokio::spawn({
+            let tx = tx.clone();
+            async move {
+                dispatch_stream_frame(
+                    &tx,
+                    Frame::new(7, MsgType::RequestBody, 0, Bytes::from_static(b"second")),
+                )
+                .await
+            }
+        });
+
+        assert_eq!(
+            stalled_send.await.expect("dispatch task should join"),
+            StreamDispatchStatus::TimedOut
+        );
+
+        let retained = rx
+            .recv()
+            .await
+            .expect("queued frame should still be present");
+        assert_eq!(retained.payload, Bytes::from_static(b"first"));
+    }
+
+    #[tokio::test]
+    async fn try_send_stream_error_emits_stream_error_frame() {
+        let (frame_tx, mut frame_rx) = bounded_queue::<Frame>(4);
+        try_send_stream_error(&frame_tx, 9, "proxy request body dispatch stalled");
+
+        let frame = frame_rx
+            .recv()
+            .await
+            .expect("stream error frame should enqueue");
+        assert_eq!(frame.stream_id, 9);
+        assert_eq!(frame.msg_type, MsgType::StreamError);
+        assert_eq!(
+            frame.payload,
+            Bytes::from_static(b"proxy request body dispatch stalled")
+        );
+    }
+}

apps/aether-proxy/src/tunnel/mod.rs

@@ -9,7 +9,7 @@ use std::sync::Arc;
 use std::time::{Duration, Instant, SystemTime, UNIX_EPOCH};
 
 use tokio::sync::watch;
-use tracing::{error, info};
+use tracing::{debug, error, info};
 
 use crate::state::{AppState, ServerContext};
 
@@ -83,7 +83,7 @@ pub async fn run(
                 return;
             }
             Ok(client::TunnelOutcome::Disconnected) => {
-                info!(server = %server.server_label, conn = conn_idx, "tunnel disconnected, reconnecting");
+                debug!(server = %server.server_label, conn = conn_idx, "tunnel disconnected, reconnecting");
             }
             Err(e) => {
                 error!(server = %server.server_label, conn = conn_idx, error = %e, "tunnel connection error, reconnecting");
@@ -114,13 +114,23 @@ pub async fn run(
             consecutive_failures,
             reconnect_salt,
         );
-        info!(
-            server = %server.server_label,
-            conn = conn_idx,
-            failures = consecutive_failures,
-            delay_ms = reconnect_delay.as_millis(),
-            "waiting before reconnect"
-        );
+        if reconnect_delay.is_zero() && consecutive_failures <= 1 {
+            debug!(
+                server = %server.server_label,
+                conn = conn_idx,
+                failures = consecutive_failures,
+                delay_ms = reconnect_delay.as_millis(),
+                "waiting before reconnect"
+            );
+        } else {
+            info!(
+                server = %server.server_label,
+                conn = conn_idx,
+                failures = consecutive_failures,
+                delay_ms = reconnect_delay.as_millis(),
+                "waiting before reconnect"
+            );
+        }
 
         tokio::select! {
             _ = tokio::time::sleep(reconnect_delay) => {}

apps/aether-proxy/src/tunnel/stream_handler.rs

@@ -16,7 +16,7 @@ use futures_util::StreamExt;
 use http_body_util::BodyExt;
 use hyper::body::Frame as BodyFrame;
 use tokio::sync::mpsc;
-use tracing::{debug, warn};
+use tracing::{debug, info, warn};
 
 use crate::state::{AppState, ServerContext};
 use crate::target_filter;
@@ -116,6 +116,110 @@ struct UpstreamResponseContext {
     request_timing: upstream_client::RequestTiming,
 }
 
+#[derive(Clone, Copy)]
+struct StreamLogContext<'a> {
+    server: &'a ServerContext,
+    stream_id: u32,
+    method: &'a hyper::Method,
+    url: Option<&'a url::Url>,
+    redirect_count: usize,
+    request_body_size: usize,
+}
+
+fn parse_request_method(method: &str) -> hyper::Method {
+    method.parse().unwrap_or(hyper::Method::GET)
+}
+
+fn request_log_host(url: &url::Url) -> &str {
+    url.host_str().unwrap_or("")
+}
+
+fn request_log_port(url: &url::Url) -> u16 {
+    url.port_or_known_default().unwrap_or(0)
+}
+
+fn request_log_path(url: &url::Url) -> &str {
+    let path = url.path();
+    if path.is_empty() {
+        "/"
+    } else {
+        path
+    }
+}
+
+fn stream_log_context<'a>(
+    server: &'a ServerContext,
+    stream_id: u32,
+    method: &'a hyper::Method,
+    url: Option<&'a url::Url>,
+    redirect_count: usize,
+    request_body_size: usize,
+) -> StreamLogContext<'a> {
+    StreamLogContext {
+        server,
+        stream_id,
+        method,
+        url,
+        redirect_count,
+        request_body_size,
+    }
+}
+
+fn log_stream_success(ctx: StreamLogContext<'_>, status: u16, duration: Duration) {
+    let url = ctx
+        .url
+        .expect("successful requests should always have a URL");
+    info!(
+        server = %ctx.server.server_label,
+        stream_id = ctx.stream_id,
+        method = %ctx.method,
+        scheme = url.scheme(),
+        host = request_log_host(url),
+        port = request_log_port(url),
+        path = request_log_path(url),
+        query_present = url.query().is_some(),
+        status,
+        duration_ms = duration.as_millis() as u64,
+        redirect_count = ctx.redirect_count,
+        request_body_bytes = ctx.request_body_size,
+        "proxy request completed"
+    );
+}
+
+fn log_stream_failure(ctx: StreamLogContext<'_>, error: &str, duration: Duration) {
+    match ctx.url {
+        Some(url) => {
+            warn!(
+                server = %ctx.server.server_label,
+                stream_id = ctx.stream_id,
+                method = %ctx.method,
+                scheme = url.scheme(),
+                host = request_log_host(url),
+                port = request_log_port(url),
+                path = request_log_path(url),
+                query_present = url.query().is_some(),
+                error = %error,
+                duration_ms = duration.as_millis() as u64,
+                redirect_count = ctx.redirect_count,
+                request_body_bytes = ctx.request_body_size,
+                "proxy request failed"
+            );
+        }
+        None => {
+            warn!(
+                server = %ctx.server.server_label,
+                stream_id = ctx.stream_id,
+                method = %ctx.method,
+                error = %error,
+                duration_ms = duration.as_millis() as u64,
+                redirect_count = ctx.redirect_count,
+                request_body_bytes = ctx.request_body_size,
+                "proxy request failed"
+            );
+        }
+    }
+}
+
 impl PreparedRequestBody {
     fn streaming(
         body_rx: mpsc::Receiver<TunnelFrame>,
@@ -466,6 +570,8 @@ async fn execute_upstream_request(
 async fn relay_upstream_response(
     server: &ServerContext,
     stream_id: u32,
+    method: &hyper::Method,
+    request_url: &url::Url,
     frame_tx: &FrameSender,
     response: hyper::Response<hyper::body::Incoming>,
     total_dns_ms: u64,
@@ -516,6 +622,18 @@ async fn relay_upstream_response(
     )
     .await
     {
+        log_stream_failure(
+            stream_log_context(
+                server,
+                stream_id,
+                method,
+                Some(request_url),
+                redirect_count,
+                request_body_size.load(Ordering::Relaxed),
+            ),
+            "tunnel response headers relay failed",
+            total_elapsed,
+        );
         return Some(total_elapsed);
     }
 
@@ -531,6 +649,18 @@ async fn relay_upstream_response(
                     )
                     .await
                     {
+                        log_stream_failure(
+                            stream_log_context(
+                                server,
+                                stream_id,
+                                method,
+                                Some(request_url),
+                                redirect_count,
+                                request_body_size.load(Ordering::Relaxed),
+                            ),
+                            "tunnel response body relay failed",
+                            total_elapsed,
+                        );
                         return Some(total_elapsed);
                     }
                 } else {
@@ -550,6 +680,18 @@ async fn relay_upstream_response(
                         )
                         .await
                         {
+                            log_stream_failure(
+                                stream_log_context(
+                                    server,
+                                    stream_id,
+                                    method,
+                                    Some(request_url),
+                                    redirect_count,
+                                    request_body_size.load(Ordering::Relaxed),
+                                ),
+                                "tunnel response body relay failed",
+                                total_elapsed,
+                            );
                             return Some(total_elapsed);
                         }
                         offset = end;
@@ -559,13 +701,26 @@ async fn relay_upstream_response(
             Err(error) => {
                 server.metrics.stream_errors.fetch_add(1, Ordering::Release);
                 warn!(stream_id, error = %error, "upstream body read error");
+                let error_message = format!("upstream body read error: {error}");
+                log_stream_failure(
+                    stream_log_context(
+                        server,
+                        stream_id,
+                        method,
+                        Some(request_url),
+                        redirect_count,
+                        request_body_size.load(Ordering::Relaxed),
+                    ),
+                    &error_message,
+                    total_elapsed,
+                );
                 send_error(frame_tx, stream_id, &format!("body read error: {error}")).await;
                 return Some(total_elapsed);
             }
         }
     }
 
-    let _ = send_frame(
+    if !send_frame(
         frame_tx,
         TunnelFrame::new(
             stream_id,
@@ -574,7 +729,22 @@ async fn relay_upstream_response(
             Bytes::new(),
         ),
     )
-    .await;
+    .await
+    {
+        log_stream_failure(
+            stream_log_context(
+                server,
+                stream_id,
+                method,
+                Some(request_url),
+                redirect_count,
+                request_body_size.load(Ordering::Relaxed),
+            ),
+            "tunnel stream end relay failed",
+            total_elapsed,
+        );
+        return Some(total_elapsed);
+    }
 
     debug!(
         stream_id,
@@ -582,6 +752,18 @@ async fn relay_upstream_response(
         redirects = redirect_count,
         "stream completed"
     );
+    log_stream_success(
+        stream_log_context(
+            server,
+            stream_id,
+            method,
+            Some(request_url),
+            redirect_count,
+            request_body_size.load(Ordering::Relaxed),
+        ),
+        status,
+        total_elapsed,
+    );
     Some(total_elapsed)
 }
 
@@ -606,6 +788,8 @@ pub async fn handle_stream(
     body_rx: mpsc::Receiver<TunnelFrame>,
     frame_tx: FrameSender,
 ) {
+    let request_method = parse_request_method(&meta.method);
+    let request_url = url::Url::parse(&meta.url).ok();
     let permit = match state.try_acquire_stream_permit().await {
         Ok(permit) => permit,
         Err(err) => {
@@ -615,6 +799,18 @@ pub async fn handle_stream(
                     "proxy admission unavailable"
                 }
             };
+            log_stream_failure(
+                stream_log_context(
+                    &server,
+                    stream_id,
+                    &request_method,
+                    request_url.as_ref(),
+                    0,
+                    0,
+                ),
+                message,
+                Duration::ZERO,
+            );
             send_error(&frame_tx, stream_id, message).await;
             return;
         }
@@ -660,9 +856,15 @@ async fn handle_stream_inner(
     body_rx: mpsc::Receiver<TunnelFrame>,
     frame_tx: &FrameSender,
 ) -> Option<Duration> {
+    let mut current_method: hyper::Method = parse_request_method(&meta.method);
     let mut current_url = match url::Url::parse(&meta.url) {
         Ok(u) => u,
         Err(e) => {
+            log_stream_failure(
+                stream_log_context(server, stream_id, &current_method, None, 0, 0),
+                &format!("invalid URL: {e}"),
+                Duration::ZERO,
+            );
             send_error(frame_tx, stream_id, &format!("invalid URL: {e}")).await;
             return None;
         }
@@ -672,12 +874,13 @@ async fn handle_stream_inner(
     match current_url.scheme() {
         "http" | "https" => {}
         other => {
-            send_error(
-                frame_tx,
-                stream_id,
-                &format!("unsupported URL scheme: {other}"),
-            )
-            .await;
+            let error_message = format!("unsupported URL scheme: {other}");
+            log_stream_failure(
+                stream_log_context(server, stream_id, &current_method, Some(&current_url), 0, 0),
+                &error_message,
+                Duration::ZERO,
+            );
+            send_error(frame_tx, stream_id, &error_message).await;
             return None;
         }
     }
@@ -685,7 +888,6 @@ async fn handle_stream_inner(
     let deadline = Instant::now()
         + Duration::from_secs(meta.timeout.clamp(MIN_TIMEOUT_SECS, MAX_TIMEOUT_SECS));
     let follow_redirects = follow_redirects_enabled(&meta);
-    let mut current_method: hyper::Method = meta.method.parse().unwrap_or(hyper::Method::GET);
     let mut current_headers = sanitize_upstream_headers(&meta.headers);
     let timeout = Duration::from_secs(meta.timeout.clamp(MIN_TIMEOUT_SECS, MAX_TIMEOUT_SECS));
     let request_body_size = Arc::new(AtomicUsize::new(0));
@@ -700,6 +902,18 @@ async fn handle_stream_inner(
         {
             Ok(body) => body,
             Err(message) => {
+                log_stream_failure(
+                    stream_log_context(
+                        server,
+                        stream_id,
+                        &current_method,
+                        Some(&current_url),
+                        0,
+                        request_body_size.load(Ordering::Relaxed),
+                    ),
+                    &message,
+                    Duration::ZERO,
+                );
                 send_error(frame_tx, stream_id, &message).await;
                 return None;
             }
@@ -715,6 +929,18 @@ async fn handle_stream_inner(
 
     loop {
         let Some(remaining) = remaining_timeout(deadline) else {
+            log_stream_failure(
+                stream_log_context(
+                    server,
+                    stream_id,
+                    &current_method,
+                    Some(&current_url),
+                    redirects_followed,
+                    request_body_size.load(Ordering::Relaxed),
+                ),
+                "upstream timeout",
+                overall_start.elapsed(),
+            );
             send_error(frame_tx, stream_id, "upstream timeout").await;
             return None;
         };
@@ -722,12 +948,20 @@ async fn handle_stream_inner(
             Some(mode) => match prepared_body.build_redirect_request_body(mode) {
                 Some(body) => body,
                 None => {
-                    send_error(
-                        frame_tx,
-                        stream_id,
-                        "upstream redirect error: body not replayable",
-                    )
-                    .await;
+                    let error_message = "upstream redirect error: body not replayable";
+                    log_stream_failure(
+                        stream_log_context(
+                            server,
+                            stream_id,
+                            &current_method,
+                            Some(&current_url),
+                            redirects_followed,
+                            request_body_size.load(Ordering::Relaxed),
+                        ),
+                        error_message,
+                        overall_start.elapsed(),
+                    );
+                    send_error(frame_tx, stream_id, error_message).await;
                     return None;
                 }
             },
@@ -748,6 +982,18 @@ async fn handle_stream_inner(
         {
             Ok(context) => context,
             Err(message) => {
+                log_stream_failure(
+                    stream_log_context(
+                        server,
+                        stream_id,
+                        &current_method,
+                        Some(&current_url),
+                        redirects_followed,
+                        request_body_size.load(Ordering::Relaxed),
+                    ),
+                    &message,
+                    overall_start.elapsed(),
+                );
                 send_error(frame_tx, stream_id, &message).await;
                 return None;
             }
@@ -767,6 +1013,8 @@ async fn handle_stream_inner(
                     return relay_upstream_response(
                         server,
                         stream_id,
+                        &current_method,
+                        &current_url,
                         frame_tx,
                         response_ctx.response,
                         total_dns_ms,
@@ -791,12 +1039,20 @@ async fn handle_stream_inner(
                     continue;
                 }
                 RedirectDecision::Error(message) => {
-                    send_error(
-                        frame_tx,
-                        stream_id,
-                        &format!("upstream redirect error: {message}"),
-                    )
-                    .await;
+                    let error_message = format!("upstream redirect error: {message}");
+                    log_stream_failure(
+                        stream_log_context(
+                            server,
+                            stream_id,
+                            &current_method,
+                            Some(&current_url),
+                            redirects_followed,
+                            request_body_size.load(Ordering::Relaxed),
+                        ),
+                        &error_message,
+                        overall_start.elapsed(),
+                    );
+                    send_error(frame_tx, stream_id, &error_message).await;
                     return None;
                 }
             }
@@ -805,6 +1061,8 @@ async fn handle_stream_inner(
         return relay_upstream_response(
             server,
             stream_id,
+            &current_method,
+            &current_url,
             frame_tx,
             response_ctx.response,
             total_dns_ms,