Rebuild: one advisory lock per group when writing releases, not per key

A picture group's batch has 30-55k release keys; an advisory lock each
ran Postgres out of shared memory (max_locks_per_transaction), and those
batches failed with HTTP 500 (the ranges went back, nothing stored).
A chunk only matches releases of its range's group, so one lock per
group serializes exactly the writers that could both start an upload.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

app/services/fragment_merger.ts

@@ -89,7 +89,7 @@ class FragmentMerger {
     }
     await ChunkPartitions.ensure()
     await db.transaction(async (trx) => {
-      // In key order, as batches lock keys (no deadlocks).
+      // In key order (batches lock their group; the merger only ever waits on one batch at a time).
       releases.sort((a, b) => a.chunk.key.localeCompare(b.chunk.key) || a.chunk.upload - b.chunk.upload)
       for (const { rangeId, chunk, candidates } of releases) await BatchStore.writeReleases(trx, rangeId, 0, [chunk], candidates)
       const done = [...completed.flatMap((c) => c.ids), ...incomplete.flatMap((c) => c.ids)]