modal 1.5.1.dev1__tar.gz → 1.5.1.dev3__tar.gz

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: modal
-Version: 1.5.1.dev1
+Version: 1.5.1.dev3
 Summary: Python client library for Modal
 Author-email: Modal Labs <support@modal.com>
 License-Expression: Apache-2.0

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/modal/_functions.py

@@ -661,6 +661,7 @@ class _Function(typing.Generic[P, ReturnType, OriginalReturnType], _Object, type
         min_containers: int | None = None,
         max_containers: int | None = None,
         buffer_containers: int | None = None,
+        scaleup_window: int | None = None,
         scaledown_window: int | None = None,
         max_concurrent_inputs: int | None = None,
         target_concurrent_inputs: int | None = None,
@@ -766,6 +767,7 @@ class _Function(typing.Generic[P, ReturnType, OriginalReturnType], _Object, type
             min_containers=min_containers,
             max_containers=max_containers,
             buffer_containers=buffer_containers,
+            scaleup_window=scaleup_window,
             scaledown_window=scaledown_window,
         )
 
@@ -1157,6 +1159,7 @@ class _Function(typing.Generic[P, ReturnType, OriginalReturnType], _Object, type
         min_containers: int | None = None,
         max_containers: int | None = None,
         buffer_containers: int | None = None,
+        scaleup_window: int | None = None,
         scaledown_window: int | None = None,
         target_concurrency: int | None = None,
     ) -> None:
@@ -1164,6 +1167,7 @@ class _Function(typing.Generic[P, ReturnType, OriginalReturnType], _Object, type
             min_containers=min_containers,
             max_containers=max_containers,
             buffer_containers=buffer_containers,
+            scaleup_window=scaleup_window,
             scaledown_window=scaledown_window,
             target_concurrency=target_concurrency,
         )

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/modal/_server.py

@@ -31,7 +31,10 @@ def validate_http_server_config(
         raise InvalidError("The `startup_timeout` argument must be positive.")
     if exit_grace_period is not None and exit_grace_period < 0:
         raise InvalidError("The `exit_grace_period` argument must be non-negative.")
-    if not is_server and exit_grace_period is not None and exit_grace_period > 25:
+    if is_server:
+        if exit_grace_period is not None and exit_grace_period > 3600:
+            raise InvalidError("The `exit_grace_period` argument must not exceed 3600 seconds (1 hour).")
+    elif exit_grace_period is not None and exit_grace_period > 25:
         raise InvalidError("The `exit_grace_period` argument must not exceed 25 seconds.")
 
     if not proxy_regions or not proxy_regions[0]:
@@ -101,6 +104,7 @@ class _Server:
         min_containers: int | None = None,
         max_containers: int | None = None,
         buffer_containers: int | None = None,
+        scaleup_window: int | None = None,
         scaledown_window: int | None = None,
         target_concurrency: int | None = None,
     ) -> None:
@@ -118,6 +122,9 @@ class _Server:
             # Limit this Server to avoid spinning up more than 5 containers
             server.update_autoscaler(max_containers=5)
 
+            # Require 30 seconds of sustained demand before scaling up
+            server.update_autoscaler(scaleup_window=30)
+
             # Adjust Server autoscaling to target 20 concurrent requests per replica
             server.update_autoscaler(target_concurrency=20)
 
@@ -129,6 +136,7 @@ class _Server:
         return await self._get_service_function()._update_autoscaler(
             min_containers=min_containers,
             max_containers=max_containers,
+            scaleup_window=scaleup_window,
             scaledown_window=scaledown_window,
             buffer_containers=buffer_containers,
             target_concurrency=target_concurrency,

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/modal/app.py

@@ -1284,6 +1284,7 @@ class _App:
         min_containers: int | None = None,  # Minimum number of containers to keep warm
         max_containers: int | None = None,  # Maximum number of containers
         buffer_containers: int | None = None,  # Additional idle containers under active load
+        scaleup_window: int | None = None,  # Stabilization window (seconds) of sustained demand before scaling up
         scaledown_window: int | None = None,  # Max idle time before scaling down (seconds)
         proxy: _Proxy | None = None,  # Modal Proxy to use in front of this server
         port: int = 8000,  # Port the HTTP server listens on
@@ -1326,6 +1327,7 @@ class _App:
             min_containers: Minimum number of containers to keep warm.
             max_containers: Maximum number of containers.
             buffer_containers: Additional idle containers under active load.
+            scaleup_window: Stabilization window (seconds) of sustained demand before scaling up, defaults to 1 second.
             scaledown_window: Max idle time before scaling down (seconds).
             proxy: Modal Proxy to use in front of this server.
             port: Port the HTTP server listens on.
@@ -1371,6 +1373,9 @@ class _App:
             if not isinstance(target_concurrency, int) or target_concurrency < 0:
                 raise InvalidError("The `target_concurrency` argument must be a non-negative integer.")
 
+        if scaleup_window is not None and scaleup_window <= 0:
+            raise InvalidError("`scaleup_window` must be > 0")
+
         http_config = api_pb2.HTTPConfig(
             port=port,
             proxy_regions=[routing_region],
@@ -1420,6 +1425,7 @@ class _App:
                 min_containers=min_containers,
                 max_containers=max_containers,
                 buffer_containers=buffer_containers,
+                scaleup_window=scaleup_window,
                 scaledown_window=scaledown_window,
                 proxy=proxy,
                 retries=None,  # No support for Server level retries

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/modal/app.pyi

@@ -747,6 +747,7 @@ class _App:
         min_containers: typing.Optional[int] = None,
         max_containers: typing.Optional[int] = None,
         buffer_containers: typing.Optional[int] = None,
+        scaleup_window: typing.Optional[int] = None,
         scaledown_window: typing.Optional[int] = None,
         proxy: typing.Optional[modal.proxy._Proxy] = None,
         port: int = 8000,
@@ -789,6 +790,7 @@ class _App:
             min_containers: Minimum number of containers to keep warm.
             max_containers: Maximum number of containers.
             buffer_containers: Additional idle containers under active load.
+            scaleup_window: Stabilization window (seconds) of sustained demand before scaling up, defaults to 1 second.
             scaledown_window: Max idle time before scaling down (seconds).
             proxy: Modal Proxy to use in front of this server.
             port: Port the HTTP server listens on.
@@ -1745,6 +1747,7 @@ class App:
         min_containers: typing.Optional[int] = None,
         max_containers: typing.Optional[int] = None,
         buffer_containers: typing.Optional[int] = None,
+        scaleup_window: typing.Optional[int] = None,
         scaledown_window: typing.Optional[int] = None,
         proxy: typing.Optional[modal.proxy.Proxy] = None,
         port: int = 8000,
@@ -1785,6 +1788,7 @@ class App:
             min_containers: Minimum number of containers to keep warm.
             max_containers: Maximum number of containers.
             buffer_containers: Additional idle containers under active load.
+            scaleup_window: Stabilization window (seconds) of sustained demand before scaling up, defaults to 1 second.
             scaledown_window: Max idle time before scaling down (seconds).
             proxy: Modal Proxy to use in front of this server.
             port: Port the HTTP server listens on.

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/modal/cli/shell.py

@@ -78,7 +78,7 @@ def _is_valid_modal_id(ref: str, prefix: str) -> bool:
 def _parse_sandbox_container_ref(ref: str) -> tuple[str, str | None]:
     """Parse a sandbox:container reference into (sandbox_id, container_name).
 
-    Supports the pattern `sb-xyz:container_name` for accessing sandbox containers.
+    Supports the pattern `sb-xyz:container_name` for accessing sandbox sidecars.
     Returns (ref, None) if no container name is specified.
     """
     if _is_valid_modal_id(ref, "sb-"):

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/modal/client.pyi

@@ -35,7 +35,7 @@ class _Client:
         server_url: str,
         client_type: int,
         credentials: typing.Optional[tuple[str, str]],
-        version: str = "1.5.1.dev1",
+        version: str = "1.5.1.dev3",
     ):
         """mdmd:hidden
         The Modal client object is not intended to be instantiated directly by users.
@@ -205,7 +205,7 @@ class Client:
         server_url: str,
         client_type: int,
         credentials: typing.Optional[tuple[str, str]],
-        version: str = "1.5.1.dev1",
+        version: str = "1.5.1.dev3",
     ):
         """mdmd:hidden
         The Modal client object is not intended to be instantiated directly by users.

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/modal/functions.pyi

@@ -90,6 +90,7 @@ class Function(
         min_containers: typing.Optional[int] = None,
         max_containers: typing.Optional[int] = None,
         buffer_containers: typing.Optional[int] = None,
+        scaleup_window: typing.Optional[int] = None,
         scaledown_window: typing.Optional[int] = None,
         max_concurrent_inputs: typing.Optional[int] = None,
         target_concurrent_inputs: typing.Optional[int] = None,
@@ -129,6 +130,7 @@ class Function(
             min_containers: typing.Optional[int] = None,
             max_containers: typing.Optional[int] = None,
             buffer_containers: typing.Optional[int] = None,
+            scaleup_window: typing.Optional[int] = None,
             scaledown_window: typing.Optional[int] = None,
             target_concurrency: typing.Optional[int] = None,
         ) -> None: ...
@@ -139,6 +141,7 @@ class Function(
             min_containers: typing.Optional[int] = None,
             max_containers: typing.Optional[int] = None,
             buffer_containers: typing.Optional[int] = None,
+            scaleup_window: typing.Optional[int] = None,
             scaledown_window: typing.Optional[int] = None,
             target_concurrency: typing.Optional[int] = None,
         ) -> None: ...

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/modal/server.pyi

@@ -52,6 +52,7 @@ class Server:
             min_containers: typing.Optional[int] = None,
             max_containers: typing.Optional[int] = None,
             buffer_containers: typing.Optional[int] = None,
+            scaleup_window: typing.Optional[int] = None,
             scaledown_window: typing.Optional[int] = None,
             target_concurrency: typing.Optional[int] = None,
         ) -> None:
@@ -69,6 +70,9 @@ class Server:
                 # Limit this Server to avoid spinning up more than 5 containers
                 server.update_autoscaler(max_containers=5)
 
+                # Require 30 seconds of sustained demand before scaling up
+                server.update_autoscaler(scaleup_window=30)
+
                 # Adjust Server autoscaling to target 20 concurrent requests per replica
                 server.update_autoscaler(target_concurrency=20)
 
@@ -85,6 +89,7 @@ class Server:
             min_containers: typing.Optional[int] = None,
             max_containers: typing.Optional[int] = None,
             buffer_containers: typing.Optional[int] = None,
+            scaleup_window: typing.Optional[int] = None,
             scaledown_window: typing.Optional[int] = None,
             target_concurrency: typing.Optional[int] = None,
         ) -> None:
@@ -102,6 +107,9 @@ class Server:
                 # Limit this Server to avoid spinning up more than 5 containers
                 server.update_autoscaler(max_containers=5)
 
+                # Require 30 seconds of sustained demand before scaling up
+                server.update_autoscaler(scaleup_window=30)
+
                 # Adjust Server autoscaling to target 20 concurrent requests per replica
                 server.update_autoscaler(target_concurrency=20)
 

{modal-1.5.1.dev1 → modal-1.5.1.dev3}/modal.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: modal
-Version: 1.5.1.dev1
+Version: 1.5.1.dev3
 Summary: Python client library for Modal
 Author-email: Modal Labs <support@modal.com>
 License-Expression: Apache-2.0