workbench 0.8.366__tar.gz → 0.8.367__tar.gz

{workbench-0.8.366 → workbench-0.8.367}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: workbench
-Version: 0.8.366
+Version: 0.8.367
 Summary: Workbench: A Dashboard and Python API for creating and deploying AWS SageMaker Model Pipelines
 Author-email: SuperCowPowers LLC <support@supercowpowers.com>
 License: MIT License

{workbench-0.8.366 → workbench-0.8.367}/applications/aws_dashboard/Dockerfile

@@ -29,7 +29,7 @@ RUN --mount=type=cache,target=/root/.cache/uv \
         -r /tmp/requirements.lock
 
 # Layer 2: Workbench only (changes often, ~20 MB)
-ARG WORKBENCH_VERSION=0.8.366
+ARG WORKBENCH_VERSION=0.8.367
 RUN pip install --no-deps "workbench==${WORKBENCH_VERSION}"
 
 # Copy app code and configuration (after pip installs to avoid cache invalidation)

{workbench-0.8.366 → workbench-0.8.367}/applications/compound_explorer/Dockerfile

@@ -28,7 +28,7 @@ RUN --mount=type=cache,target=/root/.cache/uv \
         -r /tmp/requirements.lock
 
 # Layer 2: Workbench only (changes often, ~20 MB)
-ARG WORKBENCH_VERSION=0.8.366
+ARG WORKBENCH_VERSION=0.8.367
 RUN pip install --no-deps "workbench==${WORKBENCH_VERSION}"
 
 # Copy app code and configuration (after pip installs to avoid cache invalidation)

{workbench-0.8.366 → workbench-0.8.367}/sagemaker_images/base/inference/Dockerfile

@@ -36,7 +36,7 @@ COPY --from=builder /usr/local/bin /usr/local/bin
 # workbench-bridges==0.2.10 retained for backward compat with existing
 # deployed-endpoint model bundles; drop in a future image rev once no
 # production endpoint references workbench_bridges.
-ARG WORKBENCH_VERSION=0.8.366
+ARG WORKBENCH_VERSION=0.8.367
 RUN pip install --no-deps "workbench==${WORKBENCH_VERSION}" && \
     pip install --no-deps "workbench-bridges==0.2.10"
 

{workbench-0.8.366 → workbench-0.8.367}/sagemaker_images/base/training/Dockerfile

@@ -35,6 +35,6 @@ COPY --from=builder /usr/local/bin /usr/local/bin
 # workbench-bridges==0.2.10 is retained for backward compatibility with
 # existing deployed-endpoint model bundles; drop in a future image rev once
 # no production endpoint references workbench_bridges.
-ARG WORKBENCH_VERSION=0.8.366
+ARG WORKBENCH_VERSION=0.8.367
 RUN pip install --no-deps "workbench==${WORKBENCH_VERSION}" && \
     pip install --no-deps "workbench-bridges==0.2.10"

{workbench-0.8.366 → workbench-0.8.367}/sagemaker_images/ml_pipelines/Dockerfile

@@ -48,7 +48,7 @@ COPY --from=builder /usr/local/bin /usr/local/bin
 # Layer 2: Workbench + Bridges (changes often, small layer)
 # workbench-bridges==0.2.10 pinned for reproducibility while we still depend
 # on it; on track for deprecation once nothing references workbench_bridges.
-ARG WORKBENCH_VERSION=0.8.366
+ARG WORKBENCH_VERSION=0.8.367
 RUN pip install --no-deps "workbench==${WORKBENCH_VERSION}" && \
     pip install --no-deps "workbench-bridges==0.2.10"
 

{workbench-0.8.366 → workbench-0.8.367}/sagemaker_images/pytorch_chem/inference/Dockerfile

@@ -40,7 +40,7 @@ COPY --from=builder /usr/local/bin /usr/local/bin
 # Layer 2: Workbench + bridges — rebuilds per version bump, ~20 MB delta.
 # workbench-bridges==0.2.10 retained for backward compat with existing
 # deployed-endpoint model bundles.
-ARG WORKBENCH_VERSION=0.8.366
+ARG WORKBENCH_VERSION=0.8.367
 RUN pip install --no-deps "workbench==${WORKBENCH_VERSION}" && \
     pip install --no-deps "workbench-bridges==0.2.10"
 

{workbench-0.8.366 → workbench-0.8.367}/sagemaker_images/pytorch_chem/training/Dockerfile

@@ -34,6 +34,6 @@ RUN --mount=type=cache,target=/root/.cache/uv \
 # Layer 2: Workbench + bridges — rebuilds per version bump, ~20 MB delta.
 # workbench-bridges==0.2.10 retained for backward compat with existing
 # deployed-endpoint model bundles.
-ARG WORKBENCH_VERSION=0.8.366
+ARG WORKBENCH_VERSION=0.8.367
 RUN pip install --no-deps "workbench==${WORKBENCH_VERSION}" && \
     pip install --no-deps "workbench-bridges==0.2.10"

{workbench-0.8.366 → workbench-0.8.367}/src/workbench/core/cloud_platform/aws/aws_account_clamp.py

@@ -1,6 +1,7 @@
 """AWSAccountClamp provides logic/functionality over a set of AWS IAM Services"""
 
 import boto3
+from botocore.config import Config
 from botocore.exceptions import (
     ClientError,
     UnauthorizedSSOTokenError,
@@ -14,6 +15,17 @@ from sagemaker.core.helper.session_helper import Session as SageSession
 from workbench.core.cloud_platform.aws.aws_session import AWSSession
 from workbench.utils.config_manager import ConfigManager, FatalConfigError
 
+# Hardened config for the SageMaker client: long-running training jobs poll
+# DescribeTrainingJob for hours, and under heavy concurrent Batch load a single
+# poll can hit a transient HttpTimeoutException. Adaptive retries + a longer read
+# timeout + a larger connection pool keep a blip from killing the whole pipeline.
+SAGEMAKER_CLIENT_CONFIG = Config(
+    connect_timeout=30,
+    read_timeout=120,
+    max_pool_connections=50,
+    retries={"max_attempts": 10, "mode": "adaptive"},
+)
+
 
 class AWSAccountClamp:
     """AWSAccountClamp: Singleton class for connecting to an AWS Account"""
@@ -132,7 +144,10 @@ class AWSAccountClamp:
         Returns:
             SageSession: A SageMaker session object
         """
-        return SageSession(boto_session=self.boto3_session)
+        return SageSession(
+            boto_session=self.boto3_session,
+            sagemaker_client=self.boto3_session.client("sagemaker", config=SAGEMAKER_CLIENT_CONFIG),
+        )
 
     def sagemaker_client(self) -> BaseClient:
         """Create a workbench SageMaker client (using our boto3 refreshable session)
@@ -140,7 +155,7 @@ class AWSAccountClamp:
         Returns:
             BaseClient: A SageMaker client object
         """
-        return self.boto3_session.client("sagemaker")
+        return self.boto3_session.client("sagemaker", config=SAGEMAKER_CLIENT_CONFIG)
 
     def check_tag_permissions(self):
         """Check if current role has permission to add tags to SageMaker endpoints.

{workbench-0.8.366 → workbench-0.8.367}/src/workbench.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: workbench
-Version: 0.8.366
+Version: 0.8.367
 Summary: Workbench: A Dashboard and Python API for creating and deploying AWS SageMaker Model Pipelines
 Author-email: SuperCowPowers LLC <support@supercowpowers.com>
 License: MIT License