fix(redis-streams): survive blocking-read timeout on idle streams

redis-py raises redis.exceptions.TimeoutError when a blocking XREADGROUP
returns no data within block_ms. On idle streams (US market closed → no
market:bars / signals:generated / trades:executed) every blocking read
times out; the unhandled exception tore down each worker's asyncio.TaskGroup
and exited the process, putting signal-generator, trade-executor and
learning-engine into CrashLoopBackOff. Catch it and keep polling.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

shared/redis_streams.py

@@ -5,6 +5,7 @@ import logging
 from typing import AsyncIterator
 
 from redis.asyncio import Redis
+from redis.exceptions import TimeoutError as RedisTimeoutError
 
 logger = logging.getLogger(__name__)
 
@@ -57,13 +58,18 @@ class StreamConsumer:
         """
         await self.ensure_group()
         while True:
-            messages = await self.redis.xreadgroup(
-                self.group,
-                self.consumer,
-                {self.stream: ">"},
-                count=batch_size,
-                block=block_ms,
-            )
+            try:
+                messages = await self.redis.xreadgroup(
+                    self.group,
+                    self.consumer,
+                    {self.stream: ">"},
+                    count=batch_size,
+                    block=block_ms,
+                )
+            except RedisTimeoutError:
+                # redis-py raises this when a blocking read returns no data
+                # within block_ms (idle stream). Expected — keep polling.
+                continue
             for _stream_name, entries in messages:
                 for msg_id, fields in entries:
                     data = json.loads(fields[b"data"])