fix(shutdown): exec uvicorn as PID 1 in Docker, and bound the graceful-shutdown wait

Two defects, both invisible until you ask the app to stop.

Docker never shut down gracefully at all. CMD ["sh","-c","uvicorn ..."] left
the shell as PID 1 with uvicorn as its child, and dash does not forward
signals, so docker stop SIGTERMed the shell and uvicorn never heard about it.
Measured on the shipped image: the full 10s grace period, exit 137, and no
"Shutting down" line in the log. Every stop, restart and image update was a
hard kill -- no WAL checkpoint, no MQTT disconnect, no virtual-printer
teardown. `exec` makes uvicorn PID 1; the rebuilt image now stops in 1s with
exit 0 and checkpoints the WAL.

Separately, uvicorn's timeout_graceful_shutdown defaults to None -- wait
forever for in-flight requests. An MJPEG camera stream is a response that
never completes (httptools' connection shutdown() only flips keep_alive on an
in-flight cycle, it never closes the transport), so one open camera tile
pinned the process until systemd SIGKILLed at 90s. The ordering makes it
unfixable from inside the app: uvicorn fires the lifespan shutdown -- the code
that tears the streams down -- only after connections drain.

All six launchers now pass --timeout-graceful-shutdown 5: Dockerfile,
deploy/bambuddy.service, the systemd unit and launchd plist from
install/install.sh, the SpoolBuddy installer's unit, and the Windows NSSM
registration. On timeout uvicorn cancels the request tasks; the camera
generators already unwind cleanly on CancelledError.

TimeoutStopSec raised to 30s on the units and stop_grace_period: 30s added to
compose, as backstops rather than the mechanism. On Windows NSSM's default
1500ms AppStopMethodConsole was force-killing uvicorn mid-teardown; raised to
15s, with the WM_CLOSE and thread-message stages skipped (uvicorn is a console
app with neither a window nor a message loop).

installers/windows/service/install-service.bat

@@ -30,7 +30,11 @@ REM "service not found" returns non-zero and we want to proceed.
 REM Register the service. NSSM wraps uvicorn so Windows treats it as a
 REM proper service (autostart, recovery, supervised restart).
 REM --loop asyncio required: uvloop can truncate VP FTP uploads (#1896).
-"%NSSM%" install Bambuddy "%PYTHON%" "-m uvicorn backend.app.main:app --host 0.0.0.0 --port %PORT% --loop asyncio"
+REM --timeout-graceful-shutdown required: uvicorn otherwise waits forever for
+REM in-flight requests, and an MJPEG camera stream is a response that never
+REM completes — one open camera tile hangs the stop until NSSM force-kills,
+REM skipping the WAL checkpoint and the MQTT / virtual-printer teardown.
+"%NSSM%" install Bambuddy "%PYTHON%" "-m uvicorn backend.app.main:app --host 0.0.0.0 --port %PORT% --loop asyncio --timeout-graceful-shutdown 5"
 if errorlevel 1 (
     echo [install-service] nssm install failed
     exit /b 1
@@ -42,6 +46,19 @@ REM Service configuration
 "%NSSM%" set Bambuddy Description "Bambuddy — local-first Bambu Lab printer manager"
 "%NSSM%" set Bambuddy Start SERVICE_AUTO_START
 
+REM Shutdown behaviour. NSSM's stop sequence is Ctrl-C, then WM_CLOSE, then a
+REM thread message, then TerminateProcess — each with a 1500 ms default wait.
+REM Uvicorn shuts down on the Ctrl-C, but needs longer than 1.5 seconds to
+REM finish: it drains in-flight requests (bounded at 5s by the flag above) and
+REM then runs the app teardown — WAL checkpoint, MQTT disconnect, virtual-
+REM printer stop. At the default timeout Windows force-killed it mid-teardown.
+REM
+REM Skip=6 drops the WM_CLOSE (2) and thread-message (4) methods: uvicorn is a
+REM console app with no window and no message loop, so both were only burning
+REM another 3 seconds before the kill. Ctrl-C is the one that works.
+"%NSSM%" set Bambuddy AppStopMethodSkip 6
+"%NSSM%" set Bambuddy AppStopMethodConsole 15000
+
 REM Environment: point DATA_DIR + LOG_DIR at ProgramData, prepend our
 REM bin/ to PATH so ffmpeg/ffprobe are found by the shutil.which() lookup
 REM in backend/app/services/layer_timelapse.py.