diff --git a/nucliadb/src/nucliadb/common/maindb/pg.py b/nucliadb/src/nucliadb/common/maindb/pg.py
index a8e5b7f3e7..ac67a5965c 100644
--- a/nucliadb/src/nucliadb/common/maindb/pg.py
+++ b/nucliadb/src/nucliadb/common/maindb/pg.py
@@ -275,9 +275,10 @@ async def initialize(self):
 
     async def finalize(self):
         async with self._lock:
-            await self.pool.close()
-            self.initialized = False
-            self.metrics_task.cancel()
+            if self.initialized:
+                await self.pool.close()
+                self.initialized = False
+                self.metrics_task.cancel()
 
     async def _report_metrics_task(self):
         while True:
diff --git a/nucliadb/tests/ndbfixtures/maindb.py b/nucliadb/tests/ndbfixtures/maindb.py
index a84418e6cc..a90d358bcb 100644
--- a/nucliadb/tests/ndbfixtures/maindb.py
+++ b/nucliadb/tests/ndbfixtures/maindb.py
@@ -106,13 +106,21 @@ async def maindb_driver(request: FixtureRequest) -> AsyncIterator[Driver]:
 async def pg_maindb_settings(pg):
     url = f"postgresql://postgres:postgres@{pg[0]}:{pg[1]}/postgres"
 
+    # We want to be sure schema migrations are always run. As some tests use
+    # this fixture and create their own driver, we need to create one here and
+    # run the migrations, so the maindb_settings fixture can still be generic
+    # and pg migrations are run
     driver = PGDriver(url=url, connection_pool_min_size=2, connection_pool_max_size=2)
     await driver.initialize()
     await run_pg_schema_migrations(driver)
+    await driver.finalize()
 
-    return DriverSettings(
+    yield DriverSettings(
         driver=DriverConfig.PG,
         driver_pg_url=url,
+        driver_pg_connection_pool_min_size=10,
+        driver_pg_connection_pool_max_size=10,
+        driver_pg_connection_pool_acquire_timeout_ms=200,
     )
 
 
@@ -129,7 +137,12 @@ async def pg_maindb_driver(pg_maindb_settings: DriverSettings):
             await cur.execute("DROP table IF EXISTS resources")
             await cur.execute("DROP table IF EXISTS catalog")
 
-        driver = PGDriver(url=url)
+        driver = PGDriver(
+            url=url,
+            connection_pool_min_size=pg_maindb_settings.driver_pg_connection_pool_min_size,
+            connection_pool_max_size=pg_maindb_settings.driver_pg_connection_pool_max_size,
+            acquire_timeout_ms=pg_maindb_settings.driver_pg_connection_pool_acquire_timeout_ms,
+        )
         await driver.initialize()
         await run_pg_schema_migrations(driver)
 
diff --git a/nucliadb/tests/ndbfixtures/utils.py b/nucliadb/tests/ndbfixtures/utils.py
index 8eb5abd555..f360344d36 100644
--- a/nucliadb/tests/ndbfixtures/utils.py
+++ b/nucliadb/tests/ndbfixtures/utils.py
@@ -22,7 +22,7 @@
 from typing import Callable, Optional
 
 from fastapi import FastAPI
-from httpx import AsyncClient
+from httpx import ASGITransport, AsyncClient
 
 from nucliadb.search import API_PREFIX
 
@@ -41,7 +41,8 @@ def _make_client_fixture(
         if root is False:
             client_base_url = f"{client_base_url}/{API_PREFIX}/v{version}"
 
-        client = AsyncClient(app=application, base_url=client_base_url)
+        transport = ASGITransport(app=application)  # type: ignore
+        client = AsyncClient(transport=transport, base_url=client_base_url)
         client.headers["X-NUCLIADB-ROLES"] = ";".join([role.value for role in roles])
         client.headers["X-NUCLIADB-USER"] = user
 
diff --git a/nucliadb/tests/reader/benchmarks/test_benchmarks_resource.py b/nucliadb/tests/reader/benchmarks/test_benchmarks_resource.py
index 1db0c8592d..66b035290f 100644
--- a/nucliadb/tests/reader/benchmarks/test_benchmarks_resource.py
+++ b/nucliadb/tests/reader/benchmarks/test_benchmarks_resource.py
@@ -23,6 +23,7 @@
 from httpx import AsyncClient
 
 from nucliadb.ingest.orm.resource import Resource
+from nucliadb.reader.api.v1.router import KB_PREFIX, RESOURCE_PREFIX
 from nucliadb_utils.tests.asyncbenchmark import AsyncBenchmarkFixture
 
 
@@ -38,36 +39,32 @@
 @pytest.mark.deploy_modes("component")
 async def test_get_resource_all(
     nucliadb_reader: AsyncClient,
-    test_resource: Resource,
+    full_resource: Resource,
     asyncbenchmark: AsyncBenchmarkFixture,
 ) -> None:
-    # resource = full_resource
-    # kbid = resource.kb.kbid
-    # rid = resource.uuid
+    resource = full_resource
+    kbid = resource.kb.kbid
+    rid = resource.uuid
 
-    # resp = await asyncbenchmark(
-    #     nucliadb_reader.get,
-    #     f"/{KB_PREFIX}/{kbid}/{RESOURCE_PREFIX}/{rid}",
-    #     params={
-    #         "show": ["basic", "origin", "relations", "values", "extracted"],
-    #         "field_type": [
-    #             "text",
-    #             "link",
-    #             "file",
-    #             "conversation",
-    #         ],
-    #         "extracted": [
-    #             "metadata",
-    #             "vectors",
-    #             "large_metadata",
-    #             "text",
-    #             "link",
-    #             "file",
-    #         ],
-    #     },
-    # )
-    # assert resp.status_code == 200
-    async def function(a):
-        return a
-
-    assert await asyncbenchmark(function, 10) == 10
+    resp = await asyncbenchmark(
+        nucliadb_reader.get,
+        f"/{KB_PREFIX}/{kbid}/{RESOURCE_PREFIX}/{rid}",
+        params={
+            "show": ["basic", "origin", "relations", "values", "extracted"],
+            "field_type": [
+                "text",
+                "link",
+                "file",
+                "conversation",
+            ],
+            "extracted": [
+                "metadata",
+                "vectors",
+                "large_metadata",
+                "text",
+                "link",
+                "file",
+            ],
+        },
+    )
+    assert resp.status_code == 200
diff --git a/nucliadb_utils/src/nucliadb_utils/audit/basic.py b/nucliadb_utils/src/nucliadb_utils/audit/basic.py
index ac102abbb8..8096adbeb0 100644
--- a/nucliadb_utils/src/nucliadb_utils/audit/basic.py
+++ b/nucliadb_utils/src/nucliadb_utils/audit/basic.py
@@ -54,7 +54,7 @@ async def report_and_send(
     ):
         logger.debug(f"AUDIT {audit_type} {kbid} {user} {origin} {rid} {audit_fields}")
 
-    async def visited(
+    def visited(
         self,
         kbid: str,
         uuid: str,