fix: reduce SQLITE_BUSY errors under concurrent load

- Increase busy_timeout from 5s to 15s
- Set synchronous=NORMAL (safe with WAL, reduces fsync)
- Limit MaxOpenConns to 4 to reduce write lock contention
- Explicit wal_autocheckpoint=1000

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

internal/storage/sqlite.go

@@ -39,10 +39,19 @@ func New(ctx context.Context, dataDir string) (*DB, error) {
 	// Configure SQLite pragmas
 	pragmas := []string{
 		"PRAGMA journal_mode=WAL",
-		"PRAGMA busy_timeout=5000",
+		"PRAGMA busy_timeout=15000",
 		"PRAGMA foreign_keys=ON",
+		"PRAGMA synchronous=NORMAL",
+		"PRAGMA wal_autocheckpoint=1000",
 	}
 
+	// Limit connection pool to reduce write contention.
+	// SQLite allows one writer at a time; multiple connections competing
+	// for the write lock cause SQLITE_BUSY errors. Keeping MaxOpenConns
+	// low reduces lock contention while still allowing concurrent reads.
+	db.SetMaxOpenConns(4)
+	db.SetMaxIdleConns(2)
+
 	for _, pragma := range pragmas {
 		if _, err := db.ExecContext(ctx, pragma); err != nil {
 			db.Close()

internal/storage/sqlite_test.go

@@ -68,8 +68,8 @@ func TestNew(t *testing.T) {
 			if err != nil {
 				t.Fatalf("failed to query busy_timeout: %v", err)
 			}
-			if timeout != 5000 {
-				t.Errorf("busy_timeout = %d, want 5000", timeout)
+			if timeout != 15000 {
+				t.Errorf("busy_timeout = %d, want 15000", timeout)
 			}
 
 			// Verify database is usable