diff --git a/scripts/benchmark_durations.py b/scripts/benchmark_durations.py
index 781d0d2d1e..7f05d72a03 100755
--- a/scripts/benchmark_durations.py
+++ b/scripts/benchmark_durations.py
@@ -81,7 +81,6 @@ FALLBACK_DURATION = {
     "test_runner/performance/test_seqscans.py::test_seqscans[vanilla-100000-100-0]": 0.55,
     "test_runner/performance/test_seqscans.py::test_seqscans[vanilla-10000000-1-0]": 12.189,
     "test_runner/performance/test_seqscans.py::test_seqscans[vanilla-10000000-1-4]": 13.899,
-    "test_runner/performance/test_startup.py::test_startup": 890.114,
     "test_runner/performance/test_startup.py::test_startup_simple": 2.51,
     "test_runner/performance/test_wal_backpressure.py::test_heavy_write_workload[neon_off-10-5-5]": 527.245,
     "test_runner/performance/test_wal_backpressure.py::test_heavy_write_workload[neon_on-10-5-5]": 583.46,
diff --git a/test_runner/performance/test_startup.py b/test_runner/performance/test_startup.py
index 366b9b0e68..301078d984 100644
--- a/test_runner/performance/test_startup.py
+++ b/test_runner/performance/test_startup.py
@@ -1,6 +1,3 @@
-from contextlib import closing
-
-import pytest
 import requests
 from fixtures.benchmark_fixture import MetricReport, NeonBenchmarker
 from fixtures.neon_fixtures import NeonEnvBuilder
@@ -81,49 +78,3 @@ def test_startup_simple(neon_env_builder: NeonEnvBuilder, zenbenchmark: NeonBenc
 
         # Imitate optimizations that console would do for the second start
         endpoint.respec(skip_pg_catalog_updates=True)
-
-
-# This test sometimes runs for longer than the global 5 minute timeout.
-@pytest.mark.timeout(900)
-def test_startup(neon_env_builder: NeonEnvBuilder, zenbenchmark: NeonBenchmarker):
-    neon_env_builder.num_safekeepers = 3
-    env = neon_env_builder.init_start()
-
-    # Start
-    env.neon_cli.create_branch("test_startup")
-    with zenbenchmark.record_duration("startup_time"):
-        endpoint = env.endpoints.create_start("test_startup")
-        endpoint.safe_psql("select 1;")
-
-    # Restart
-    endpoint.stop_and_destroy()
-    with zenbenchmark.record_duration("restart_time"):
-        endpoint.create_start("test_startup")
-        endpoint.safe_psql("select 1;")
-
-    # Fill up
-    num_rows = 1000000  # 30 MB
-    num_tables = 100
-    with closing(endpoint.connect()) as conn:
-        with conn.cursor() as cur:
-            for i in range(num_tables):
-                cur.execute(f"create table t_{i} (i integer);")
-                cur.execute(f"insert into t_{i} values (generate_series(1,{num_rows}));")
-
-    # Read
-    with zenbenchmark.record_duration("read_time"):
-        endpoint.safe_psql("select * from t_0;")
-
-    # Read again
-    with zenbenchmark.record_duration("second_read_time"):
-        endpoint.safe_psql("select * from t_0;")
-
-    # Restart
-    endpoint.stop_and_destroy()
-    with zenbenchmark.record_duration("restart_with_data"):
-        endpoint.create_start("test_startup")
-        endpoint.safe_psql("select 1;")
-
-    # Read
-    with zenbenchmark.record_duration("read_after_restart"):
-        endpoint.safe_psql("select * from t_0;")