Jason Barto jpbarto

Machine Learning 101 on SageMaker Immersion Day

Following the steps online access your AWS account through AWS Event Engine:

Visit https://github.com/awslabs/amazon-sagemaker-examples, at the bottom of the page you will find a link to create an Amazon SageMaker notebook
Follow the instructions using an ml.m5.2xlarge instance type.
When the notebook has been created click Open JupyterLab.

Clone the Amazon SageMaker Examples to your notebook from GitHub.

	#!/usr/bin/bash

	# This script is built to compile Qt5 on an Amazon Linux2-based Amazon Workspace
	# This script is known to fail on a Standard Workspace with 2 vCPU and 4 GB of RAM (Out of Memory error suspected)
	# Script has been tested on a Power Workspace with 4 vCPU and 32 GB of RAM
	# based upon documentation at
	# https://doc.qt.io/qt-5/linux-requirements.html

	set -e

	#!/usr/bin/env bash

	###############################################################################################
	#
	# The following shell script uses the AWS CLI to forcefully detach a non-root volume from a
	# running EC2 instance. In preliminary testing this produced a read-only filesystem on a
	# running host with the volume mounted. This is intended to grossly simulate communication
	# errors with the EBS volume from an EC2 instance.
	#
	# Note that this script will NOT detach the root volume from an EC2 instance.

	[
	{
	"Entry": "/public/research",
	"Target": "/maxld-public-bucket"
	},
	{
	"Entry": "/subscribed/2018/indices",
	"Target": "/maxld-subscribe-bucket/historical/2018/indices"
	},
	{

	#!/bin/bash

	set -e

	AWS_REGION=eu-west-1

	aws rds describe-db-instance-automated-backups --region $AWS_REGION --query 'DBInstanceAutomatedBackups[*].{DatabaseID:DBInstanceIdentifier,Earliest:RestoreWindow.EarliestTime,Latest:RestoreWindow.LatestTime}' --output table

	#!/bin/bash

	set -e

	AWS_REGION=eu-west-2
	SOURCE_DATABASE_NAME=database-1
	RESTORE_TIME='2020-10-28T22:02:01+00:00'

	NEW_DATABASE_NAME="rds-db-$RANDOM"

	#!/bin/bash

	# The following shell script uses Amazon Athena to query AWS CloudTrail logs for any occurrences of the
	# action sagemaker:ListNotebookInstances. The query returns the user identity who invoked the API, along
	# with the region where the API was called. A simple count of the number of invokes is outputted as a
	# result.

	SQL="SELECT useridentity.arn, eventname, eventsource, awsregion, sourceipaddress, errorcode, eventtime FROM cloudtrail_logs_account_logging WHERE eventsource = 'sagemaker.amazonaws.com' AND eventname in ('ListNotebookInstances') AND eventtime > '2020-04-15' LIMIT 3;"

	EXEC_ID=$(aws athena start-query-execution --query-string "$SQL" --result-configuration OutputLocation=s3://my-s3-logging-bucket/queries --query 'QueryExecutionId' --output text)