How to Run a Sagemaker Notebook From AWS Lambdas、、

Requirement: trigger the SageMaker notebook .ipynb file by event

Using Sagemaker LifeCycle Configurations

Create the LifeCycle Configurations with script for Start Instance. It is required Sagemaker instance has to be in Stop State before invocation

Lambda Code

Create Lambda function is Python 3.8 runtime with 60s timeout
Lambda code:

import json
import boto3

def lambda_handler(event, context):

    client = boto3.client('sagemaker')

    #wish to get current status of instance
    status = client.describe_notebook_instance(NotebookInstanceName='demo-notebook')
    print('NotebookInstanceStatus %s' % (status['NotebookInstanceStatus']))

    #Start the instance
    client.start_notebook_instance(NotebookInstanceName='demo-notebook')
    print("instance starting")

    # waiter = client.get_waiter('notebook_instance_in_service')
    # waiter.wait(
    #     NotebookInstanceName='demo-notebook',
    #     WaiterConfig={
    #         'Delay': 30,
    #         'MaxAttempts': 10
    #     }
    # )
    
    status = client.describe_notebook_instance(NotebookInstanceName='demo-notebook')
    print('NotebookInstanceStatus %s' % (status['NotebookInstanceStatus']))

2021-04-22T00:39:11.085+08:00   Successfully installed cycler-0.10.0 kiwisolver-1.3.1 matplotlib-3.3.4 pillow-8.2.0

2021-04-22T00:39:13.086+08:00   [NbConvertApp] WARNING | Config option `kernel_spec_manager_class` not recognized by `NbConvertApp`.

2021-04-22T00:39:13.086+08:00   [NbConvertApp] Converting notebook /home/ec2-user/SageMaker/notebook1.ipynb to html

2021-04-22T00:39:14.086+08:00   [NbConvertApp] Executing notebook with kernel: python3

2021-04-22T00:39:22.088+08:00   [NbConvertApp] Writing 344427 bytes to /home/ec2-user/SageMaker/notebook1.html