Backport PR jupyter-server#910 on branch 1.x (Add back support for kernel launch timeout pad) (jupyter-server#911)

Co-authored-by: Ciprian Anton <[email protected]>

Author: meeseeksmachine

jupyter_server/gateway/gateway_client.py

@@ -128,7 +128,7 @@ def connect_timeout_default(self):
             os.environ.get("JUPYTER_GATEWAY_CONNECT_TIMEOUT", self.connect_timeout_default_value)
         )
 
-    request_timeout_default_value = 40.0
+    request_timeout_default_value = 42.0
     request_timeout_env = "JUPYTER_GATEWAY_REQUEST_TIMEOUT"
     request_timeout = Float(
         default_value=request_timeout_default_value,
@@ -341,6 +341,25 @@ def gateway_retry_max_default(self):
             os.environ.get("JUPYTER_GATEWAY_RETRY_MAX", self.gateway_retry_max_default_value)
         )
 
+    launch_timeout_pad_default_value = 2.0
+    launch_timeout_pad_env = "JUPYTER_GATEWAY_LAUNCH_TIMEOUT_PAD"
+    launch_timeout_pad = Float(
+        default_value=launch_timeout_pad_default_value,
+        config=True,
+        help="""Timeout pad to be ensured between KERNEL_LAUNCH_TIMEOUT and request_timeout
+        such that request_timeout >= KERNEL_LAUNCH_TIMEOUT + launch_timeout_pad.
+        (JUPYTER_GATEWAY_LAUNCH_TIMEOUT_PAD env var)""",
+    )
+
+    @default("launch_timeout_pad")
+    def launch_timeout_pad_default(self):
+        return float(
+            os.environ.get(
+                "JUPYTER_GATEWAY_LAUNCH_TIMEOUT_PAD",
+                self.launch_timeout_pad_default_value,
+            )
+        )
+
     @property
     def gateway_enabled(self):
         return bool(self.url is not None and len(self.url) > 0)
@@ -353,12 +372,18 @@ def init_static_args(self):
         perform this operation once.
 
         """
-        # Ensure that request timeout and KERNEL_LAUNCH_TIMEOUT are the same, taking the
-        #  greater value of the two.
-        if self.request_timeout < float(GatewayClient.KERNEL_LAUNCH_TIMEOUT):
-            self.request_timeout = float(GatewayClient.KERNEL_LAUNCH_TIMEOUT)
-        elif self.request_timeout > float(GatewayClient.KERNEL_LAUNCH_TIMEOUT):
-            GatewayClient.KERNEL_LAUNCH_TIMEOUT = int(self.request_timeout)
+        # Ensure that request timeout and KERNEL_LAUNCH_TIMEOUT are in sync, taking the
+        #  greater value of the two and taking into account the following relation:
+        #  request_timeout = KERNEL_LAUNCH_TIME + padding
+        minimum_request_timeout = (
+            float(GatewayClient.KERNEL_LAUNCH_TIMEOUT) + self.launch_timeout_pad
+        )
+        if self.request_timeout < minimum_request_timeout:
+            self.request_timeout = minimum_request_timeout
+        elif self.request_timeout > minimum_request_timeout:
+            GatewayClient.KERNEL_LAUNCH_TIMEOUT = int(
+                self.request_timeout - self.launch_timeout_pad
+            )
         # Ensure any adjustments are reflected in env.
         os.environ["KERNEL_LAUNCH_TIMEOUT"] = str(GatewayClient.KERNEL_LAUNCH_TIMEOUT)
 

tests/test_gateway.py

@@ -186,6 +186,7 @@ def init_gateway(monkeypatch):
     monkeypatch.setenv("JUPYTER_GATEWAY_HTTP_USER", mock_http_user)
     monkeypatch.setenv("JUPYTER_GATEWAY_REQUEST_TIMEOUT", "44.4")
     monkeypatch.setenv("JUPYTER_GATEWAY_CONNECT_TIMEOUT", "44.4")
+    monkeypatch.setenv("JUPYTER_GATEWAY_LAUNCH_TIMEOUT_PAD", "1.1")
     yield
     GatewayClient.clear_instance()
 
@@ -198,11 +199,10 @@ async def test_gateway_env_options(init_gateway, jp_serverapp):
         jp_serverapp.gateway_config.connect_timeout == jp_serverapp.gateway_config.request_timeout
     )
     assert jp_serverapp.gateway_config.connect_timeout == 44.4
+    assert jp_serverapp.gateway_config.launch_timeout_pad == 1.1
 
     GatewayClient.instance().init_static_args()
-    assert GatewayClient.instance().KERNEL_LAUNCH_TIMEOUT == int(
-        jp_serverapp.gateway_config.request_timeout
-    )
+    assert GatewayClient.instance().KERNEL_LAUNCH_TIMEOUT == 43
 
 
 async def test_gateway_cli_options(jp_configurable_serverapp):
@@ -211,6 +211,7 @@ async def test_gateway_cli_options(jp_configurable_serverapp):
         "--GatewayClient.http_user=" + mock_http_user,
         "--GatewayClient.connect_timeout=44.4",
         "--GatewayClient.request_timeout=96.0",
+        "--GatewayClient.launch_timeout_pad=5.1",
     ]
 
     GatewayClient.clear_instance()
@@ -221,10 +222,40 @@ async def test_gateway_cli_options(jp_configurable_serverapp):
     assert app.gateway_config.http_user == mock_http_user
     assert app.gateway_config.connect_timeout == 44.4
     assert app.gateway_config.request_timeout == 96.0
+    assert app.gateway_config.launch_timeout_pad == 5.1
     GatewayClient.instance().init_static_args()
     assert (
-        GatewayClient.instance().KERNEL_LAUNCH_TIMEOUT == 96
-    )  # Ensure KLT gets set from request-timeout
+        GatewayClient.instance().KERNEL_LAUNCH_TIMEOUT == 90
+    )  # Ensure KLT gets set from request-timeout - launch_timeout_pad
+    GatewayClient.clear_instance()
+
+
+@pytest.mark.parametrize(
+    "request_timeout,kernel_launch_timeout,expected_request_timeout,expected_kernel_launch_timeout",
+    [(50, 10, 50, 45), (10, 50, 55, 50)],
+)
+async def test_gateway_request_timeout_pad_option(
+    jp_configurable_serverapp,
+    monkeypatch,
+    request_timeout,
+    kernel_launch_timeout,
+    expected_request_timeout,
+    expected_kernel_launch_timeout,
+):
+    argv = [
+        f"--GatewayClient.request_timeout={request_timeout}",
+        "--GatewayClient.launch_timeout_pad=5",
+    ]
+
+    GatewayClient.clear_instance()
+    app = jp_configurable_serverapp(argv=argv)
+
+    monkeypatch.setattr(GatewayClient, "KERNEL_LAUNCH_TIMEOUT", kernel_launch_timeout)
+    GatewayClient.instance().init_static_args()
+
+    assert app.gateway_config.request_timeout == expected_request_timeout
+    assert GatewayClient.instance().KERNEL_LAUNCH_TIMEOUT == expected_kernel_launch_timeout
+
     GatewayClient.clear_instance()