fix: enable SQLite WAL mode to prevent database locking errors (backport #4048) (#4226)

Fixes race condition causing "database is locked" errors during
concurrent writes to SQLite, particularly in streaming responses with
guardrails where multiple inference calls write simultaneously.

Enable Write-Ahead Logging (WAL) mode for SQLite which allows multiple
concurrent readers and one writer without blocking. Set busy_timeout to
5s so SQLite retries instead of failing immediately. Remove the logic
that disabled write queues for SQLite since WAL mode eliminates the
locking issues that prompted disabling them.

Fixes: test_output_safety_guardrails_safe_content[stream=True]
flake<hr>This is an automatic backport of pull request #4048 done by
[Mergify](https://mergify.com).

Signed-off-by: Charlie Doern <cdoern@redhat.com>
Co-authored-by: Ashwin Bharambe <ashwin.bharambe@gmail.com>

llama_stack/providers/utils/sqlstore/sqlalchemy_sqlstore.py

@@ -17,6 +17,7 @@ from sqlalchemy import (
     String,
     Table,
     Text,
+    event,
     inspect,
     select,
     text,
@@ -75,7 +76,36 @@ class SqlAlchemySqlStoreImpl(SqlStore):
         self.metadata = MetaData()
 
     def create_engine(self) -> AsyncEngine:
-        return create_async_engine(self.config.engine_str, pool_pre_ping=True)
+        # Configure connection args for better concurrency support
+        connect_args = {}
+        if "sqlite" in self.config.engine_str:
+            # SQLite-specific optimizations for concurrent access
+            # With WAL mode, most locks resolve in milliseconds, but allow up to 5s for edge cases
+            connect_args["timeout"] = 5.0
+            connect_args["check_same_thread"] = False  # Allow usage across asyncio tasks
+
+        engine = create_async_engine(
+            self.config.engine_str,
+            pool_pre_ping=True,
+            connect_args=connect_args,
+        )
+
+        # Enable WAL mode for SQLite to support concurrent readers and writers
+        if "sqlite" in self.config.engine_str:
+
+            @event.listens_for(engine.sync_engine, "connect")
+            def set_sqlite_pragma(dbapi_conn, connection_record):
+                cursor = dbapi_conn.cursor()
+                # Enable Write-Ahead Logging for better concurrency
+                cursor.execute("PRAGMA journal_mode=WAL")
+                # Set busy timeout to 5 seconds (retry instead of immediate failure)
+                # With WAL mode, locks should be brief; if we hit 5s there's a bigger issue
+                cursor.execute("PRAGMA busy_timeout=5000")
+                # Use NORMAL synchronous mode for better performance (still safe with WAL)
+                cursor.execute("PRAGMA synchronous=NORMAL")
+                cursor.close()
+
+        return engine
 
     async def create_table(
         self,