fix: split HTTP connect/read timeouts (Jellyfin build + qBit stats)

Every HTTP client passed an integer timeout to requests, applying the same
value to BOTH connect and read phases. A slow Jellyfin /Items page or qBit
/sync/maindata blew through the 10s read budget → ReadTimeoutError. Split
into a (connect=5s, read=60s default) tuple via shared http_timeout() helper.
The media index build worker uses a 180s read floor. Existing services with
low timeout_seconds benefit from bumping to 60+.

CHANGELOG.md

@@ -4,6 +4,19 @@ All notable changes to Manage. Breaking changes are marked with **BREAKING**.
 
 ## [Unreleased]
 
+### Fixed — HTTP read timeouts
+
+- Service HTTP clients now use a `(connect, read)` timeout tuple (connect 5s,
+  read 60s default) instead of a single integer, resolving `ReadTimeoutError`
+  on slow Jellyfin index builds and qBittorrent stats. The media index build
+  worker uses a 180s read floor so slow `/Items` pages on large libraries
+  don't time out mid-build.
+- The shared `http_timeout()` helper (`clients/http_timeout.py`) decouples
+  connect (fail-fast on dead hosts) from read (generous for slow responses).
+- Integration `timeout_seconds` defaults were raised from 5/10s to 15/60s.
+- Existing services with a low `timeout_seconds` may benefit from bumping it
+  to 60+ via the service editor.
+
 ### **BREAKING** — Prometheus queries now route through Grafana gateway
 
 - The `prometheus` service config changed: `base_url` is replaced by