Do not fail a node because the engine's own stdout is gone

The log tee wrote through to the real stream unguarded, and the worker
pool tees a returned call's logs there after reading its result and
before handing it back — so a dead stdout, which `fluksio serve` makes
possible by running the engine as a child of the dashboard holding that
pipe, failed the node with its outputs already in hand. The capture half
runs first, so swallowing the write loses nothing.

Also: `flow_events` catches the RuntimeError a peer leaving mid-send
raises, which is a disconnect by another route, and the remote agent no
longer raises out of the task when its subprocess died before it could
be written to — the read below reports that and ends the call.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TXQv6KNyyvY7Z1etYTUUAd

backend/fluksio/flow/logs.py

@@ -34,10 +34,20 @@ class _Tee(io.TextIOBase):
         sink = _sink.get()
         if sink is not None and text:
             sink(text)
-        return self._real.write(text)
+        try:
+            return self._real.write(text)
+        except OSError:
+            # A dead stdout — `fluksio serve` runs the engine as a child of the
+            # dashboard, which holds the far end of that pipe — must not fail
+            # the node whose output was being teed. The capture above has it,
+            # and there is nowhere left to report the loss to anyway.
+            return len(text)
 
     def flush(self) -> None:
-        self._real.flush()
+        try:
+            self._real.flush()
+        except OSError:
+            pass
 
     def isatty(self) -> bool:
         return self._real.isatty()