fix(sync): don't try a timed-out feed again

A request that hung for the full timeout says the way out is hanging, not
that the endpoint is flaky; three of them per channel stretched a poll over
a dead egress to about three times its interval.

backend/app/youtube/rss.py

@@ -12,8 +12,11 @@ RSS_URL = "https://www.youtube.com/feeds/videos.xml?channel_id={channel_id}"
 # The waits before a second and a third try. YouTube's feed endpoint answers the SAME channel with
 # 404, 500 and 200 in turn (measured 2026-09-19: 15 of 24 channels failed a first request, 16 of them
 # had succeeded by the fifth, one second apart), so one failure is not an answer. A feed that is
-# really gone costs two more requests per poll.
+# really gone costs two more requests per poll. A TIMEOUT is not tried again: the flaky endpoint
+# answers fast, and a request that hung for `TIMEOUT` says the way out is hanging — three of them
+# per channel would stretch a poll over a dead egress to three times its interval.
 RETRY_DELAYS = (1.0, 2.0)
+TIMEOUT = 20.0
 # A 404 is in the set because that is how the flaky endpoint fails most often; a 403/410 is not,
 # and neither is anything else that says the request itself is wrong.
 _RETRYABLE_STATUS = frozenset({404, 429, 500, 502, 503, 504})
@@ -41,11 +44,13 @@ def _get_feed(channel_id: str) -> bytes:
         if attempt:
             time.sleep(RETRY_DELAYS[attempt - 1])
         try:
-            resp = httpx.get(url, timeout=20.0, headers={"User-Agent": "Siftlode/1.0"})
+            resp = httpx.get(url, timeout=TIMEOUT, headers={"User-Agent": "Siftlode/1.0"})
         # silent-ok: not silent — the last failure is raised after the final try.
         except httpx.HTTPError as exc:
             error = RssError(f"RSS fetch failed for {channel_id}: {exc}", "transport")
             error.__cause__ = exc
+            if isinstance(exc, httpx.TimeoutException):
+                break
             continue
         if resp.status_code == 200:
             return resp.content

backend/tests/test_sync_guards.py

@@ -282,6 +282,14 @@ class TestRssErrorVsEmpty:
         assert rss.fetch_channel_feed("UC_blip")
         assert len(tries) == 2
 
+    def test_a_timeout_is_not_tried_again(self, monkeypatch):
+        # A hung request already cost the full timeout; retrying it would triple a poll over a
+        # dead egress.
+        tries, waits = self._answers(monkeypatch, httpx.ReadTimeout("hung"), 200)
+        with pytest.raises(rss.RssError) as err:
+            rss.fetch_channel_feed("UC_hung")
+        assert (len(tries), waits, err.value.reason) == (1, [], "transport")
+
     def test_a_feed_that_keeps_failing_raises_its_last_failure(self, monkeypatch):
         tries, _ = self._answers(monkeypatch, 500, 404, 404)
         with pytest.raises(rss.RssError) as err: