Amazon SageMaker endpoints and quotas - AWS General Reference

Amazon SageMaker endpoints and quotas

The following are the service endpoints and service quotas for this service. To connect programmatically to an AWS service, you use an endpoint. In addition to the standard AWS endpoints, some AWS services offer FIPS endpoints in selected Regions. For more information, see AWS service endpoints. Service quotas, also referred to as limits, are the maximum number of service resources or operations for your AWS account. For more information, see AWS service quotas.

Service Endpoints

The following table provides a list of Region-specific endpoints that SageMaker supports for training and deploying models. This include creating and managing notebook instances, training jobs, model, endpoint configurations, and endpoints.

Region Name Region Endpoint Protocol
US East (Ohio) us-east-2

api.sagemaker.us-east-2.amazonaws.com

api-fips.sagemaker.us-east-2.amazonaws.com

HTTPS

HTTPS

US East (N. Virginia) us-east-1

api.sagemaker.us-east-1.amazonaws.com

api-fips.sagemaker.us-east-1.amazonaws.com

HTTPS

HTTPS

US West (N. California) us-west-1

api.sagemaker.us-west-1.amazonaws.com

api-fips.sagemaker.us-west-1.amazonaws.com

HTTPS

HTTPS

US West (Oregon) us-west-2

api.sagemaker.us-west-2.amazonaws.com

api-fips.sagemaker.us-west-2.amazonaws.com

HTTPS

HTTPS

Africa (Cape Town) af-south-1 api.sagemaker.af-south-1.amazonaws.com HTTPS
Asia Pacific (Hong Kong) ap-east-1 api.sagemaker.ap-east-1.amazonaws.com HTTPS
Asia Pacific (Mumbai) ap-south-1 api.sagemaker.ap-south-1.amazonaws.com HTTPS
Asia Pacific (Seoul) ap-northeast-2 api.sagemaker.ap-northeast-2.amazonaws.com HTTPS
Asia Pacific (Singapore) ap-southeast-1 api.sagemaker.ap-southeast-1.amazonaws.com HTTPS
Asia Pacific (Sydney) ap-southeast-2 api.sagemaker.ap-southeast-2.amazonaws.com HTTPS
Asia Pacific (Tokyo) ap-northeast-1 api.sagemaker.ap-northeast-1.amazonaws.com HTTPS
Canada (Central) ca-central-1 api.sagemaker.ca-central-1.amazonaws.com HTTPS
China (Beijing) cn-north-1 api.sagemaker.cn-north-1.amazonaws.com.cn HTTPS
China (Ningxia) cn-northwest-1 api.sagemaker.cn-northwest-1.amazonaws.com.cn HTTPS
Europe (Frankfurt) eu-central-1 api.sagemaker.eu-central-1.amazonaws.com HTTPS
Europe (Ireland) eu-west-1 api.sagemaker.eu-west-1.amazonaws.com HTTPS
Europe (London) eu-west-2 api.sagemaker.eu-west-2.amazonaws.com HTTPS
Europe (Milan) eu-south-1 api.sagemaker.eu-south-1.amazonaws.com HTTPS
Europe (Paris) eu-west-3 api.sagemaker.eu-west-3.amazonaws.com HTTPS
Europe (Stockholm) eu-north-1 api.sagemaker.eu-north-1.amazonaws.com HTTPS
Middle East (Bahrain) me-south-1 api.sagemaker.me-south-1.amazonaws.com HTTPS
South America (São Paulo) sa-east-1 api.sagemaker.sa-east-1.amazonaws.com HTTPS
AWS GovCloud (US) us-gov-west-1

api.sagemaker.us-gov-west-1.amazonaws.com

api-fips.sagemaker.us-gov-west-1.amazonaws.com

api.sagemaker.us-gov-west-1.amazonaws.com

HTTPS

HTTPS

HTTPS

The following table provides a list of Region-specific endpoints that Amazon SageMaker supports for making inference requests against models hosted in SageMaker.

Region Name Region Endpoint Protocol
US East (Ohio) us-east-2

runtime.sagemaker.us-east-2.amazonaws.com

runtime-fips.sagemaker.us-east-2.amazonaws.com

HTTPS

HTTPS

US East (N. Virginia) us-east-1

runtime.sagemaker.us-east-1.amazonaws.com

runtime-fips.sagemaker.us-east-1.amazonaws.com

HTTPS

HTTPS

US West (N. California) us-west-1

runtime.sagemaker.us-west-1.amazonaws.com

runtime-fips.sagemaker.us-west-1.amazonaws.com

HTTPS

HTTPS

US West (Oregon) us-west-2

runtime.sagemaker.us-west-2.amazonaws.com

runtime-fips.sagemaker.us-west-2.amazonaws.com

HTTPS

HTTPS

Africa (Cape Town) af-south-1 runtime.sagemaker.af-south-1.amazonaws.com HTTPS
Asia Pacific (Hong Kong) ap-east-1 runtime.sagemaker.ap-east-1.amazonaws.com HTTPS
Asia Pacific (Mumbai) ap-south-1 runtime.sagemaker.ap-south-1.amazonaws.com HTTPS
Asia Pacific (Seoul) ap-northeast-2 runtime.sagemaker.ap-northeast-2.amazonaws.com HTTPS
Asia Pacific (Singapore) ap-southeast-1 runtime.sagemaker.ap-southeast-1.amazonaws.com HTTPS
Asia Pacific (Sydney) ap-southeast-2 runtime.sagemaker.ap-southeast-2.amazonaws.com HTTPS
Asia Pacific (Tokyo) ap-northeast-1 runtime.sagemaker.ap-northeast-1.amazonaws.com HTTPS
Canada (Central) ca-central-1 runtime.sagemaker.ca-central-1.amazonaws.com HTTPS
China (Beijing) cn-north-1 runtime.sagemaker.cn-north-1.amazonaws.com.cn HTTPS
China (Ningxia) cn-northwest-1 runtime.sagemaker.cn-northwest-1.amazonaws.com.cn HTTPS
Europe (Frankfurt) eu-central-1 runtime.sagemaker.eu-central-1.amazonaws.com HTTPS
Europe (Ireland) eu-west-1 runtime.sagemaker.eu-west-1.amazonaws.com HTTPS
Europe (London) eu-west-2 runtime.sagemaker.eu-west-2.amazonaws.com HTTPS
Europe (Milan) eu-south-1 runtime.sagemaker.eu-south-1.amazonaws.com HTTPS
Europe (Paris) eu-west-3 runtime.sagemaker.eu-west-3.amazonaws.com HTTPS
Europe (Stockholm) eu-north-1 runtime.sagemaker.eu-north-1.amazonaws.com HTTPS
Middle East (Bahrain) me-south-1 runtime.sagemaker.me-south-1.amazonaws.com HTTPS
South America (São Paulo) sa-east-1 runtime.sagemaker.sa-east-1.amazonaws.com HTTPS
AWS GovCloud (US) us-gov-west-1 runtime.sagemaker.us-gov-west-1.amazonaws.com HTTPS

Service Quotas

SageMaker quotas for new accounts might be different from the default quotas listed here. If you receive an error that you've exceeded your quota, contact customer service to request a quota increase for the resources you want to use.

On-demand and Spot instance quotas are tracked and modified separately. For example, with the default quotas, you could run up to 20 training jobs with on-demand ml.m4.xlarge instances and up to 20 training jobs with Managed Spot ml.m4.xlarge instances simultaneously. Request quota increases for on-demand and spot instances separately.

Amazon SageMaker Notebooks
Resource Default
ml.t2.medium instances 20
ml.t2.large instances 20
ml.t2.xlarge instances 20
ml.t2.2xlarge instances 20
ml.t3.medium instances 20
ml.t3.large instances 20
ml.t3.xlarge instances 20
ml.t3.2xlarge instances 20
ml.m4.xlarge instances 20
ml.m4.2xlarge instances 20
ml.m4.4xlarge instances 10
ml.m4.10xlarge instances 5
ml.m4.16xlarge instances 5
ml.m5.xlarge instances 20
ml.m5.2xlarge instances 20
ml.m5.4xlarge instances 10
ml.m5.12xlarge instances 3
ml.m5.24xlarge instances 2
ml.c4.xlarge instances 20
ml.c4.2xlarge instances 20
ml.c4.4xlarge instances 20
ml.c4.8xlarge instances 20
ml.c5.xlarge instances 20
ml.c5.2xlarge instances 20
ml.c5.4xlarge instances 5
ml.c5.9xlarge instances 5
ml.c5.18xlarge instances 5
ml.c5d.xlarge instances 20
ml.c5d.2xlarge instances 20
ml.c5d.4xlarge instances 5
ml.c5d.9xlarge instances 5
ml.c5d.18xlarge instances 5
ml.p2.xlarge instances 1
ml.p2.8xlarge instances 1
ml.p2.16xlarge instances 1
ml.p3.2xlarge instances 2
ml.p3.8xlarge instances 2
ml.p3.16xlarge instances 2
Number of notebook instances 20
Amazon SageMaker Automatic Model Tuning
Resource Default
Number of concurrent hyperparameter tuning jobs 100
Number of hyperparameters that can be searched (every possible value in a categorical hyperparameter counts against this quota) 20
Number of metrics defined per hyperparameter tuning job 20
Number of parallel training jobs per hyperparameter tuning job 10
Number of training jobs per hyperparameter tuning job 500
Maximum run time for a hyperparameter tuning job 30 days
Amazon SageMaker Processing
Resource Default
ml.c4.2xlarge 20
ml.c4.4xlarge 20
ml.c4.8xlarge 20
ml.c4.xlarge 20
ml.c5.18xlarge 5
ml.c5.2xlarge 20
ml.c5.4xlarge 5
ml.c5.9xlarge 5
ml.c5.xlarge 20
ml.m4.10xlarge 5
ml.m4.16xlarge 5
ml.m4.2xlarge 20
ml.m4.4xlarge 10
ml.m4.xlarge 20
ml.m5.12xlarge 4
ml.m5.24xlarge 4
ml.m5.2xlarge 20
ml.m5.4xlarge 10
ml.m5.large 20
ml.m5.xlarge 65
ml.p2.16xlarge 4
ml.p2.8xlarge 4
ml.p2.xlarge 4
ml.p3.16xlarge 4
ml.p3.2large 4
ml.p3.8xlarge 4
ml.r5.12xlarge 20
ml.r5.16xlarge 20
ml.r5.24xlarge 20
ml.r5.2xlarge 20
ml.r5.4xlarge 20
ml.r5.8xlarge 20
ml.r5.large 20
ml.r5.xlarge 20
ml.t3.2xlarge 5
ml.t3.large 20
ml.t3.medium 50
ml.t3.xlarge 10
Longest run time for a processing job 5 days
Number of instances across processing jobs 75
Number of instances for a processing job 20
Size of EBS volume for an instance 1 TB
Amazon SageMaker Training and Managed Spot Training
Resource Default
ml.m4.xlarge instances 20
ml.m4.2xlarge instances 20
ml.m4.4xlarge instances 10
ml.m4.10xlarge instances 5
ml.m4.16xlarge instances 5
ml.m5.large instances 20
ml.m5.xlarge instances 20
ml.m5.2xlarge instances 20
ml.m5.4xlarge instances 10
ml.m5.12xlarge instances 3
ml.m5.24xlarge instances 2
ml.c4.xlarge instances 20
ml.c4.2xlarge instances 20
ml.c4.4xlarge instances 20
ml.c4.8xlarge instances 20
ml.c5.xlarge instances 20
ml.c5.2xlarge instances 20
ml.c5.4xlarge instances 5
ml.c5.9xlarge instances 5
ml.c5.18xlarge instances 5
ml.p2.xlarge instances 1
ml.p2.8xlarge instances 1
ml.p2.16xlarge instances 1
ml.p3.2xlarge instances 2
ml.p3.8xlarge instances 2
ml.p3.16xlarge instances 2
Longest run time for a training job 5 days
Number of instances across training jobs 20
Number of instances for a training job 20
Size of EBS volume for an instance 1 TB
Amazon SageMaker Hosting
Resource Default
ml.t2.medium instances 20
ml.t2.large instances 20
ml.t2.xlarge instances 20
ml.t2.2xlarge instances 20
ml.m4.xlarge instances 20
ml.m4.2xlarge instances 20
ml.m4.4xlarge instances 10
ml.m4.10xlarge instances 5
ml.m4.16xlarge instances 5
ml.m5.large instances 20
ml.m5.xlarge instances 20
ml.m5.2xlarge instances 20
ml.m5.4xlarge instances 10
ml.m5.12xlarge instances 3
ml.m5.24xlarge instances 2
ml.m5d.large instances 20
ml.m5d.xlarge instances 20
ml.m5d.2xlarge instances 20
ml.m5d.4xlarge instances 10
ml.m5d.12xlarge instances 3
ml.m5d.24xlarge instances 2
ml.c4.large instances 20
ml.c4.xlarge instances 20
ml.c4.2xlarge instances 20
ml.c4.4xlarge instances 20
ml.c4.8xlarge instances 20
ml.c5.large instances 20
ml.c5.xlarge instances 20
ml.c5.2xlarge instances 20
ml.c5.4xlarge instances 5
ml.c5.9xlarge instances 5
ml.c5.18xlarge instances 5
ml.c5d.large instances 20
ml.c5d.xlarge instances 20
ml.c5d.2xlarge instances 20
ml.c5d.4xlarge instances 5
ml.c5d.9xlarge instances 5
ml.c5d.18xlarge instances 5
ml.p2.xlarge instances 2
ml.p2.8xlarge instances 2
ml.p2.16xlarge instances 2
ml.p3.2xlarge instances 2
ml.p3.8xlarge instances 2
ml.p3.16xlarge instances 2
ml.g4dn.xlarge instances 2
ml.g4dn.2xlarge instances 2
ml.g4dn.4xlarge instances 2
ml.g4dn.8xlarge instances 2
ml.g4dn.12xlarge instances 2
ml.g4dn.16xlarge instances 2
ml.r5.large instances 5
ml.r5.xlarge instances 5
ml.r5.2xlarge instances 4
ml.r5.4xlarge instances 4
ml.r5.12xlarge instances 3
ml.r5.24xlarge instances 3
ml.r5d.large instances 5
ml.r5d.xlarge instances 5
ml.r5d.2xlarge instances 4
ml.r5d.4xlarge instances 4
ml.r5d.12xlarge instances 3
ml.r5d.24xlarge instances 3
Number of instances across active endpoints 20
Number of instances for an endpoint 20
Total TPS for all endpoints 10,000
Maximum payload size for endpoint invocation 25 MB
Inference timeout for endpoint invocation 60 seconds
Amazon SageMaker Batch Transform
Resource Default
ml.m4.xlarge instances 20
ml.m4.2xlarge instances 20
ml.m4.4xlarge instances 10
ml.m4.10xlarge instances 5
ml.m4.16xlarge instances 5
ml.m5.large instances 20
ml.m5.xlarge instances 20
ml.m5.2xlarge instances 20
ml.m5.4xlarge instances 10
ml.m5.12xlarge instances 3
ml.m5.24xlarge instances 2
ml.c4.xlarge instances 20
ml.c4.2xlarge instances 20
ml.c4.4xlarge instances 20
ml.c4.8xlarge instances 20
ml.c5.xlarge instances 20
ml.c5.2xlarge instances 20
ml.c5.4xlarge instances 5
ml.c5.9xlarge instances 5
ml.c5.18xlarge instances 5
ml.p2.xlarge instances 1
ml.p2.8xlarge instances 1
ml.p2.16xlarge instances 1
ml.p3.2xlarge instances 2
ml.p3.8xlarge instances 2
ml.p3.16xlarge instances 2
Longest run time for a transform job 5 days
Number of instances for a transform job 20
Maximum payload size for mini-batch inference 100 MB
Inference timeout for mini-batch inference 60 minutes
Amazon SageMaker Ground Truth
Resource Default
Concurrent labeling jobs 20
Dataset objects per labeling job 100,000