fix(agentex): run SSE status recheck on a timer, not only when idle

The lost-terminal-event recheck was gated on idle cycles, so steady non-terminal
traffic (message_count stays nonzero) meant it never ran and a subscriber could
stay open after a terminal transition whose event publish failed. Drive the
recheck off its own interval timer so it fires busy or idle; the keepalive ping
stays idle-only.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: deepthi-rao-scale

agentex/src/domain/use_cases/streams_use_case.py

@@ -127,6 +127,7 @@ async def stream_task_events(
         # client's read fails on each cycle; without backoff this turns into a
         # log-ingestion firehose (one failure per client per cycle, ~once/sec).
         consecutive_errors = 0
+        last_status_check = last_message_time
         try:
             # Application-level control loop
             while True:
@@ -161,17 +162,22 @@ async def stream_task_events(
                     # healthy again, so reset the backoff/error counter.
                     consecutive_errors = 0
 
-                    # Idle: recheck status (catches a lost terminal event), else ping.
+                    current_time = asyncio.get_running_loop().time()
+                    # Authoritative status recheck on an interval, busy or idle,
+                    # in case a terminal event's publish was lost.
+                    if current_time - last_status_check >= ping_interval:
+                        last_status_check = current_time
+                        task = await self.task_service.get_task(id=task_id)
+                        if task.status in TERMINAL_TASK_STATUSES:
+                            logger.info(
+                                f"Ending SSE stream for task {task_id}: "
+                                "terminal on status recheck"
+                            )
+                            return
+
+                    # Idle: send keepalive ping so proxies don't reap us.
                     if message_count == 0:
-                        current_time = asyncio.get_running_loop().time()
                         if current_time - last_message_time >= ping_interval:
-                            task = await self.task_service.get_task(id=task_id)
-                            if task.status in TERMINAL_TASK_STATUSES:
-                                logger.info(
-                                    f"Ending SSE stream for task {task_id}: "
-                                    "terminal on status recheck"
-                                )
-                                return
                             yield ":ping\n\n"
                             last_message_time = current_time
                         await asyncio.sleep(0.1)