fix: add scaled to zero in client (#4712)

Signed-off-by: FogDong <fog@bentoml.com>
bentoml · May 7, 2024 · e19f765 · e19f765
1 parent 6de2041
commit e19f765
Showing 1 changed file with 8 additions and 2 deletions.
diff --git a/src/bentoml/_internal/cloud/deployment.py b/src/bentoml/_internal/cloud/deployment.py
@@ -453,7 +453,10 @@ def get_client(
         from _bentoml_impl.client import SyncHTTPClient
 
         self._refetch()
-        if self._schema.status != DeploymentStatus.Running.value:
+        if (
+            self._schema.status != DeploymentStatus.Running.value
+            and self._schema.status != DeploymentStatus.ScaledToZero.value
+        ):
             raise BentoMLException(f"Deployment status is {self._schema.status}")
         if self._urls is None or len(self._urls) != 1:
             raise BentoMLException("Deployment url is not ready")
@@ -468,7 +471,10 @@ def get_async_client(
         from _bentoml_impl.client import AsyncHTTPClient
 
         self._refetch()
-        if self._schema.status != DeploymentStatus.Running.value:
+        if (
+            self._schema.status != DeploymentStatus.Running.value
+            and self._schema.status != DeploymentStatus.ScaledToZero.value
+        ):
             raise BentoMLException(f"Deployment status is {self._schema.status}")
         if self._urls is None or len(self._urls) != 1:
             raise BentoMLException("Deployment url is not ready")