fix: release QA hardening across processing, media, security, and CI gates (#649)

A release-readiness QA pass over the whole product. The commits split into
defects a user would hit and gates that were reporting green while measuring
nothing.

## Fixes that change behaviour

Rate limiting was bypassable on every install: TRUST_PROXY defaulted to true, so
request.ip came from a client-set header and a forged X-Forwarded-For got past
the login limiter. The default is now a private-network trust list.

A transient Postgres outage stranded in-flight jobs, leaving finished output on
disk with no row pointing at it. A reconciler now resolves those rows and adopts
the bytes rather than dropping the work.

A Redis connection that moved to a new address wedged every read-blocked
consumer, so completions stopped signalling while health still answered 200.
Socket timeouts plus subscriber pings recover it.

Installing more than one AI bundle left the shared venv multi-versioned and
silently broke three tools. The installer now reconciles distributions to one
version each.

Converting an image to JXL at quality 1 through 4 returned a 500, because
libjxl 0.7 rejects the distance those values compute. The quality is floored at
what the encoder honours. A missing ffmpeg was also reported to the user as a
corrupt upload; it now says the engine is unavailable.

RAW uploads reached an unpatched LibRaw on arm64, so it is built from source at
0.22.2, and the release scan was split so it can fail on an unfixed critical
instead of hiding it behind ignore-unfixed.

## Gates that could not fail

Two mutation lanes ran zero mutants because Stryker crawled the gitignored docs
build; coverage discarded its whole report on any failing test; the lint gate
skipped root tests, scripts, and two workspaces; and several generated matrices
counted a host missing ffmpeg as a passing tool. Each now measures what it
claims.

Full evidence and the outstanding release items are tracked locally and are not
part of this branch.

apps/docs/guide/database.md

@@ -142,6 +142,17 @@ Security-relevant action log.
 | `details` | jsonb | Action-specific data |
 | `createdAt` | timestamp | Action time |
 
+### user_preferences {#user-preferences}
+
+Per-user UI state, keyed by preference name. Backs the dashboard's pinned tools through `PUT /api/v1/preferences`.
+
+| Column | Type | Notes |
+|---|---|---|
+| `userId` | text | FK to users, cascades on delete. Primary key with `key` |
+| `key` | text | Preference name. Primary key with `userId` |
+| `value` | jsonb | Preference payload |
+| `updatedAt` | timestamp | Last write |
+
 ## Migrations {#migrations}
 
 Drizzle handles schema migrations. Migration files live in `apps/api/drizzle/`. During development:
@@ -158,25 +169,33 @@ In production, pending migrations are applied automatically on startup.
 
 The relational database lives in the Postgres container's `SnapOtter-pgdata` volume, not the app's `/data` volume.
 
-**Option 1: pg_dump (recommended)**
+**Logical backup with validation (recommended)**
 
 ```bash
-# Dump the database while the stack is running
-docker exec SnapOtter-postgres pg_dump -U snapotter snapotter > backup.sql
+# Dump into PostgreSQL's portable custom archive format
+docker exec SnapOtter-postgres \
+  pg_dump --format=custom --no-owner -U snapotter snapotter > snapotter.dump
+test -s snapotter.dump
+docker exec -i SnapOtter-postgres pg_restore --list < snapotter.dump >/dev/null
 
-# Restore into a fresh database
-cat backup.sql | docker exec -i SnapOtter-postgres psql -U snapotter snapotter
+# Restore into a fresh/disposable target first and fail on the first SQL error
+docker exec -i SnapOtter-postgres \
+  pg_restore --exit-on-error --clean --if-exists --no-owner \
+  -U snapotter -d snapotter < snapotter.dump
 ```
 
-**Option 2: Volume snapshot**
+This database dump does not contain saved library objects in `/data/files` or durable BullMQ state in Redis. Back up and restore those with the coordinated procedure in [Security & Hardening](/guide/security#backup-and-recovery).
+
+**Cold volume snapshot**
 
 ```bash
-# Stop the stack, then snapshot the pgdata volume
-docker compose down
-docker run --rm -v SnapOtter-pgdata:/data -v $(pwd)/backup:/backup \
-  alpine tar czf /backup/snapotter-pgdata.tar.gz -C /data .
+# Stop every service first, then use your storage platform to snapshot the
+# PostgreSQL, app-data, and Redis volumes as one crash-consistent set.
+docker compose -f docker/docker-compose.yml stop
 ```
 
+Do not copy a live PostgreSQL data directory with `tar`. Compose prefixes volume names by project, so resolve the mounted volume IDs from `docker inspect` or your storage platform rather than assuming the literal label `SnapOtter-pgdata`.
+
 ### Migrating from 1.x (SQLite) {#migrating-from-1-x-sqlite}
 
 Upgrading from SnapOtter 1.x has its own guide: see [Upgrading from 1.x to 2.0](./upgrading). In short, reuse your existing `/data` volume and 2.0 auto-detects and imports `/data/snapotter.db` on first boot (or set `SQLITE_MIGRATE_PATH` to point at it explicitly). Back up the whole `/data` volume first, not just `snapotter.db`: 1.x uses SQLite WAL mode, so a stopped container often leaves most of its data in `snapotter.db-wal` beside an almost-empty `snapotter.db`.