AWS sagemaker documentation change
Summary
Restructured and updated documentation for SageMaker CloudWatch metrics. Added new 'SageMaker AI endpoint metrics' section with detailed resource utilization metrics (CPUReservation, GPUReservation, etc.), reorganized metric categories, and standardized section naming conventions.
Security assessment
The changes focus on improving metric documentation clarity and organization. While metrics like CPU/GPU reservation could aid in resource monitoring, there is no explicit evidence of addressing a security vulnerability or weakness. The note about CloudWatch metric retention periods is operational guidance, not security-related.
Diff
diff --git a/sagemaker/latest/dg/monitoring-cloudwatch.md b/sagemaker/latest/dg/monitoring-cloudwatch.md index ca6056fc0..ec038572a 100644 --- a//sagemaker/latest/dg/monitoring-cloudwatch.md +++ b//sagemaker/latest/dg/monitoring-cloudwatch.md @@ -5 +5 @@ -Endpoint Invocation MetricsSageMaker AI inference component metricsMulti-Model Endpoint MetricsJobs and Endpoint MetricsInference Recommender MetricsGround Truth MetricsFeature Store MetricsPipelines Metrics +Endpoint metricsEndpoint invocation metricsInference component metricsMulti-model endpoint metricsJob metricsInference Recommender metricsGround Truth metricsFeature Store metricsPipelines metrics @@ -14,0 +15,2 @@ To graph metrics without using a search, specify its exact name in the source vi + * SageMaker AI endpoint metrics + @@ -21 +23 @@ To graph metrics without using a search, specify its exact name in the source vi - * SageMaker AI jobs and endpoint metrics + * SageMaker AI job metrics @@ -33,0 +36,28 @@ To graph metrics without using a search, specify its exact name in the source vi +## SageMaker AI endpoint metrics + +The `/aws/sagemaker/Endpoints` namespace includes the following metrics for endpoint instances. + +Metrics are available at a 1-minute frequency. + +###### Note + +Amazon CloudWatch supports [high-resolution custom metrics](https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/publishingMetrics.html) and its finest resolution is 1 second. However, the finer the resolution, the shorter the lifespan of the CloudWatch metrics. For the 1-second frequency resolution, the CloudWatch metrics are available for 3 hours. For more information about the resolution and the lifespan of the CloudWatch metrics, see [GetMetricStatistics](https://docs.aws.amazon.com/AmazonCloudWatch/latest/APIReference/API_GetMetricStatistics.html) in the _Amazon CloudWatch API Reference_. + +Endpoint metrics Metric | Description +---|--- +`CPUReservation` | The sum of CPUs reserved by containers on an instance. This metric is provided only for endpoints that host active inference components. The value ranges between 0%–100%. In the settings for an inference component, you set the CPU reservation with the `NumberOfCpuCoresRequired` parameter. For example, if there 4 CPUs, and 2 are reserved, the `CPUReservation` metric is 50%. +`CPUUtilization` | The sum of each individual CPU core's utilization. The CPU utilization of each core range is 0–100. For example, if there are four CPUs, the `CPUUtilization` range is 0%–400%. For endpoint variants, the value is the sum of the CPU utilization of the primary and supplementary containers on the instance. Units: Percent +`CPUUtilizationNormalized` | The normalized sum of the utilization of each individual CPU core. This metric is provided only for endpoints that host active inference components. The value ranges between 0%–100%. For example, if there are four CPUs, and the `CPUUtilization` metric is 200%, then the `CPUUtilizationNormalized` metric is 50%. +`DiskUtilization` | The percentage of disk space used by the containers on an instance. This value range is 0%–100%.For endpoint variants, the value is the sum of the disk space utilization of the primary and supplementary containers on the instance.Units: Percent +`GPUMemoryUtilization` | The percentage of GPU memory used by the containers on an instance. The value range is 0–100 and is multiplied by the number of GPUs. For example, if there are four GPUs, the `GPUMemoryUtilization` range is 0%–400%. For endpoint variants, the value is the sum of the GPU memory utilization of the primary and supplementary containers on the instance. Units: Percent +`GPUMemoryUtilizationNormalized` | The normalized percentage of GPU memory used by the containers on an instance. This metric is provided only for endpoints that host active inference components. The value ranges between 0%–100%. For example, if there are four GPUs, and the `GPUMemoryUtilization` metric is 200%, then the `GPUMemoryUtilizationNormalized` metric is 50%. +`GPUReservation` | The sum of GPUs reserved by containers on an instance. This metric is provided only for endpoints that host active inference components. The value ranges between 0%–100%. In the settings for an inference component, you set the GPU reservation by `NumberOfAcceleratorDevicesRequired`. For example, if there are 4 GPUs and 2 are reserved, the `GPUReservation` metric is 50%. +`GPUUtilization` | The percentage of GPU units that are used by the containers on an instance. The value can range between 0–100 and is multiplied by the number of GPUs. For example, if there are four GPUs, the `GPUUtilization` range is 0%–400%. For endpoint variants, the value is the sum of the GPU utilization of the primary and supplementary containers on the instance. Units: Percent +`GPUUtilizationNormalized` | The normalized percentage of GPU units that are used by the containers on an instance. This metric is provided only for endpoints that host active inference components. The value ranges between 0%–100%. For example, if there are four GPUs, and the `GPUUtilization` metric is 200%, then the `GPUUtilizationNormalized` metric is 50%. +`MemoryReservation` | The sum of memory reserved by containers on an instance. This metric is provided only for endpoints that host active inference components. The value ranges between 0%–100%. In the settings for an inference component, you set the memory reservation with the `MinMemoryRequiredInMb` parameter. For example, if a 32 GiB instance reserved 1024 MB, the `MemoryReservation` metric would be 3.125%. +`MemoryUtilization` | The percentage of memory that is used by the containers on an instance. This value range is 0%–100%. For endpoint variants, the value is the sum of the memory utilization of the primary and supplementary containers on the instance. Units: Percent + +Dimensions for endpoint metrics Dimension | Description +---|--- +`EndpointName, VariantName` | Filters endpoint metrics for a `ProductionVariant` of the specified endpoint and variant. + @@ -55,3 +85 @@ For more information about total latency, see [Best practices for load testing A -**Endpoint Invocation Metrics** - -Metric | Description +Endpoint invocation metrics Metric | Description @@ -70 +97,0 @@ Metric | Description -| @@ -72,3 +99 @@ Metric | Description -**Dimensions for Endpoint Invocation Metrics** - -Dimension | Description +Dimensions for endpoint invocation metrics Dimension | Description @@ -85 +110 @@ Metrics are available at a 1-minute frequency. -Metric | Description +Inference component metrics Metric | Description @@ -92,3 +117 @@ Metric | Description -**Dimensions for Inference Component Metrics** - -Dimension | Description +Dimensions for inference component metrics Dimension | Description @@ -106,3 +129 @@ For information about how long CloudWatch metrics are retained for, see [GetMetr -**Multi-Model Endpoint Model Loading Metrics** - -Metric | Description +Multi-model endpoint model loading metrics Metric | Description @@ -116,3 +137 @@ Metric | Description -**Dimensions for Multi-Model Endpoint Model Loading Metrics** - -Dimension | Description +Dimensions for multi-model endpoint model loading metrics Dimension | Description @@ -128,3 +147 @@ For information about how long CloudWatch metrics are retained for, see [GetMetr -**Multi-Model Endpoint Model Instance Metrics** - -Metric | Description +Multi-model endpoint model instance metrics Metric | Description @@ -134,3 +151 @@ Metric | Description -**Dimensions for Multi-Model Endpoint Model Loading Metrics** - -Dimension | Description +Dimensions for multi-model endpoint model loading metrics Dimension | Description @@ -140 +155 @@ Dimension | Description -## SageMaker AI jobs and endpoint metrics +## SageMaker AI job metrics @@ -142 +157 @@ Dimension | Description -The `/aws/sagemaker/ProcessingJobs`, `/aws/sagemaker/TrainingJobs`, `/aws/sagemaker/TransformJobs`, and `/aws/sagemaker/Endpoints` namespaces include the following metrics for training jobs and endpoint instances. +The `/aws/sagemaker/ProcessingJobs`, `/aws/sagemaker/TrainingJobs`, and `/aws/sagemaker/TransformJobs` namespaces include the following metrics for processing jobs, training jobs, and batch transform jobs. @@ -154,3 +169 @@ To profile your training job with a finer resolution down to 100-millisecond (0. -**Processing Job, Training Job, Batch Transform Job, and Endpoint Instance Metrics** - -Metric | Description +Processing job, training job, and batch transform job metrics Metric | Description @@ -158,2 +171 @@ Metric | Description -`CPUReservation` | The sum of CPUs reserved by containers on an instance. The value ranges between 0%–100%. In the settings for an inference component, you set the CPU reservation with the `NumberOfCpuCoresRequired` parameter. For example, if there 4 CPUs, and 2 are reserved, the `CPUReservation` metric is 50%. -`CPUUtilization` | The sum of each individual CPU core's utilization. The CPU utilization of each core range is 0–100. For example, if there are four CPUs, the `CPUUtilization` range is 0%–400%. For processing jobs, the value is the CPU utilization of the processing container on the instance.For training jobs, the value is the CPU utilization of the algorithm container on the instance.For batch transform jobs, the value is the CPU utilization of the transform container on the instance.For endpoint variants, the value is the sum of the CPU utilization of the primary and supplementary containers on the instance. +`CPUUtilization` | The sum of each individual CPU core's utilization. The CPU utilization of each core range is 0–100. For example, if there are four CPUs, the `CPUUtilization` range is 0%–400%. For processing jobs, the value is the CPU utilization of the processing container on the instance.For training jobs, the value is the CPU utilization of the algorithm container on the instance.For batch transform jobs, the value is the CPU utilization of the transform container on the instance. @@ -164,2 +176 @@ For multi-instance jobs, each instance reports CPU utilization metrics. However, -`CPUUtilizationNormalized` | The normalized sum of the utilization of each individual CPU core. The value ranges between 0%–100%. For example, if there are four CPUs, and the `CPUUtilization` metric is 200%, then the `CPUUtilizationNormalized` metric is 50%. -`DiskUtilization` | The percentage of disk space used by the containers on an instance uses. This value range is 0%–100%. This metric is not supported for batch transform jobs.For processing jobs, the value is the disk space utilization of the processing container on the instance.For training jobs, the value is the disk space utilization of the algorithm container on the instance.For endpoint variants, the value is the sum of the disk space utilization of the primary and supplementary containers on the instance.Units: Percent +`DiskUtilization` | The percentage of disk space used by the containers on an instance. This value range is 0%–100%. This metric is not supported for batch transform jobs.For processing jobs, the value is the disk space utilization of the processing container on the instance.For training jobs, the value is the disk space utilization of the algorithm container on the instance.Units: Percent @@ -170 +181 @@ For multi-instance jobs, each instance reports disk utilization metrics. However -`GPUMemoryUtilization` | The percentage of GPU memory used by the containers on an instance. The value range is 0–100 and is multiplied by the number of GPUs. For example, if there are four GPUs, the `GPUMemoryUtilization` range is 0%–400%.For processing jobs, the value is the GPU memory utilization of the processing container on the instance.For training jobs, the value is the GPU memory utilization of the algorithm container on the instance.For batch transform jobs, the value is the GPU memory utilization of the transform container on the instance.For endpoint variants, the value is the sum of the GPU memory utilization of the primary and supplementary containers on the instance. +`GPUMemoryUtilization` | The percentage of GPU memory used by the containers on an instance. The value range is 0–100 and is multiplied by the number of GPUs. For example, if there are four GPUs, the `GPUMemoryUtilization` range is 0%–400%.For processing jobs, the value is the GPU memory utilization of the processing container on the instance.For training jobs, the value is the GPU memory utilization of the algorithm container on the instance.For batch transform jobs, the value is the GPU memory utilization of the transform container on the instance. @@ -175,3 +186 @@ For multi-instance jobs, each instance reports GPU memory utilization metrics. H -`GPUMemoryUtilizationNormalized` | The normalized percentage of GPU memory used by the containers on an instance. The value ranges between 0%–100%. For example, if there are four GPUs, and the `GPUMemoryUtilization` metric is 200%, then the `GPUMemoryUtilizationNormalized` metric is 50%. -`GPUReservation` | The sum of GPUs reserved by containers on an instance. The value ranges between 0%–100%. In the settings for an inference component, you set the GPU reservation by `NumberOfAcceleratorDevicesRequired`. For example, if there are 4 GPUs and 2 are reserved, the `GPUReservation` metric is 50%. -`GPUUtilization` | The percentage of GPU units that are used by the containers on an instance. The value can range between 0–100 and is multiplied by the number of GPUs. For example, if there are four GPUs, the `GPUUtilization` range is 0%–400%.For processing jobs, the value is the GPU utilization of the processing container on the instance.For training jobs, the value is the GPU utilization of the algorithm container on the instance.For batch transform jobs, the value is the GPU utilization of the transform container on the instance.For endpoint variants, the value is the sum of the GPU utilization of the primary and supplementary containers on the instance. +`GPUUtilization` | The percentage of GPU units that are used by the containers on an instance. The value can range between 0–100 and is multiplied by the number of GPUs. For example, if there are four GPUs, the `GPUUtilization` range is 0%–400%.For processing jobs, the value is the GPU utilization of the processing container on the instance.For training jobs, the value is the GPU utilization of the algorithm container on the instance.For batch transform jobs, the value is the GPU utilization of the transform container on the instance. @@ -182,3 +191 @@ For multi-instance jobs, each instance reports GPU utilization metrics. However, -`GPUUtilizationNormalized` | The normalized percentage of GPU units that are used by the containers on an instance. The value ranges between 0%–100%. For example, if there are four GPUs, and the `GPUUtilization` metric is 200%, then the `GPUUtilizationNormalized` metric is 50%. -`MemoryReservation` | The sum of memory reserved by containers on an instance. The value ranges between 0%–100%. In the settings for an inference component, you set the memory reservation with the `MinMemoryRequiredInMb` parameter. For example, if a 32 GiB instance reserved 1024 MB, the `MemoryReservation` metric would be 3.125%. -`MemoryUtilization` | The percentage of memory that is used by the containers on an instance. This value range is 0%–100%.For processing jobs, the value is the memory utilization of the processing container on the instance.For training jobs, the value is the memory utilization of the algorithm container on the instance.For batch transform jobs, the value is the memory utilization of the transform container on the instance.For endpoint variants, the value is the sum of the memory utilization of the primary and supplementary containers on the instance.Units: Percent +`MemoryUtilization` | The percentage of memory that is used by the containers on an instance. This value range is 0%–100%.For processing jobs, the value is the memory utilization of the processing container on the instance.For training jobs, the value is the memory utilization of the algorithm container on the instance.For batch transform jobs, the value is the memory utilization of the transform container on the instance.Units: Percent @@ -190,3 +197 @@ For multi-instance jobs, each instance reports memory utilization metrics. Howev -**Dimensions for Processing Job, Training Job and Batch Transform Job Instance Metrics** - -Dimension | Description +Dimensions for job metrics Dimension | Description @@ -200,3 +205 @@ The `/aws/sagemaker/InferenceRecommendationsJobs` namespace includes the followi -**Inference Recommender Metrics** - -Metric | Description +Inference Recommender metrics Metric | Description @@ -209,3 +212 @@ Metric | Description -**Dimensions for Inference Recommender Job Metrics** - -Dimension | Description +Dimensions for Inference Recommender job metrics Dimension | Description @@ -218,3 +219 @@ Dimension | Description -**Ground Truth Metrics** - -Metric | Description +Ground Truth metrics Metric | Description @@ -236,3 +235 @@ Metric | Description -**Dimensions for Dataset Object Metrics** - -Dimension | Description +Dimensions for dataset object metrics Dimension | Description @@ -244,3 +241 @@ Dimension | Description -**Feature Store Consumption Metrics** - -Metric | Description +Feature Store consumption metrics Metric | Description @@ -253,3 +248 @@ Metric | Description -**Dimensions for Feature Store Consumption Metrics** - -Dimension | Description +Dimensions for Feature Store consumption metrics Dimension | Description @@ -259,3 +252 @@ Dimension | Description -**Feature Store Operational Metrics** - -Metric | Description +Feature Store operational metrics Metric | Description @@ -269,3 +260 @@ Metric | Description -**Dimensions for Feature Store Operational Metrics** - -Dimension | Description +Dimensions for Feature Store operational metrics Dimension | Description @@ -280 +269 @@ The `AWS/Sagemaker/ModelBuildingPipeline` namespace includes the following metri -Two categories of Pipelines execution metrics are available: +Two categories of pipeline execution metrics are available: @@ -291,3 +280 @@ Metrics are available at a 1-minute frequency. -**Pipelines Execution Metrics** - -Metric | Description +Pipeline execution metrics Metric | Description @@ -301,3 +288 @@ Metric | Description -**Dimensions for Execution Metrics by Pipeline** - -Dimension | Description +Dimensions for pipeline execution metrics Dimension | Description @@ -307,2 +291,0 @@ Dimension | Description -**Pipelines Step Metrics** - @@ -313 +296 @@ Metrics are available at a 1-minute frequency. -Metric | Description +Pipeline step metrics Metric | Description @@ -321,3 +304 @@ Metric | Description -**Dimensions for Pipelines Step Metrics** - -Dimension | Description +Dimensions for pipeline step metrics Dimension | Description