SageMaker Endpoint
HTTPS endpoint serving an ML model for real-time inference.
Configuration
Section titled “Configuration”| Setting | Type | Required | Default |
|---|---|---|---|
| Endpoint name | Text | Yes | — |
| Endpoint config name | Text | — | — |
| Instance type Options: ml.t3.medium — 2v · 4 GiB, ml.t3.large — 2v · 8 GiB, ml.t3.xlarge — 4v · 16 GiB, ml.m5.large — 2v · 8 GiB, ml.m5.xlarge — 4v · 16 GiB, ml.m5.2xlarge — 8v · 32 GiB, ml.m5.4xlarge — 16v · 64 GiB, ml.m5.12xlarge — 48v · 192 GiB, ml.m5.24xlarge — 96v · 384 GiB, ml.c5.large — 2v · 4 GiB, ml.c5.xlarge — 4v · 8 GiB, ml.c5.4xlarge — 16v · 32 GiB, ml.c5.9xlarge — 36v · 72 GiB, ml.r5.large — 2v · 16 GiB, ml.r5.xlarge — 4v · 32 GiB, ml.r5.2xlarge — 8v · 64 GiB, ml.g4dn.xlarge — 4v · 16 GiB, ml.g4dn.2xlarge — 8v · 32 GiB, ml.g4dn.4xlarge — 16v · 64 GiB, ml.g4dn.8xlarge — 32v · 128 GiB, ml.g5.xlarge — 4v · 16 GiB, ml.g5.2xlarge — 8v · 32 GiB, ml.g5.4xlarge — 16v · 64 GiB, ml.g5.8xlarge — 32v · 128 GiB, ml.g5.12xlarge — 48v · 192 GiB, ml.g5.24xlarge — 96v · 384 GiB, ml.g5.48xlarge — 192v · 768 GiB, ml.p3.2xlarge — 8v · 61 GiB, ml.p3.8xlarge — 32v · 244 GiB, ml.p3.16xlarge — 64v · 488 GiB, ml.p4d.24xlarge — 96v · 1152 GiB | Choice | — | ml.m5.large |
| Model artifact S3 URL | Text | — | — |
| Container image (ECR URI) | Text | — | — |
| KMS key ARN (encryption) | Text | — | — |
| Container env vars | Key–value | — | — |
| Serverless inference | Toggle | — | false |
| Tags | Key–value | — | — |
Connections
Section titled “Connections”| Socket | Direction | Accepts | Terraform arg |
|---|---|---|---|
| Endpoint config | Input | aws.sagemaker-endpoint-configuration | endpoint_config_name |
| Inference client | Output | any | — |