Skip to content

Understand your AWS data from a SageMaker notebook

Install it in your notebook and get readable reports on each AWS service: what's there, what it costs, and what to do next, without leaving Jupyter.

  • One install, boto3 only
  • Read-only: never changes your data
  • Plain text outside Jupyter

Guides

Each guide covers setting up in SageMaker, every command with examples of its output, and the IAM permissions it needs.

  • s3.py

    Amazon S3

    Buckets and the files in them.

    • Every bucket's size, monthly cost and risks
    • Folder trees, and search by name, size or date
    • Preview CSV, Parquet, JSON, Excel, PDF, Word and more
    • Cut storage costs and recover deleted files

    Open the S3 guide →

  • s3_explorer.py

    S3 file explorer

    Click through buckets and folders, and see inside each file.

    • Your buckets and folders, one click at a time
    • What's inside a file, as soon as you click it
    • PDFs page by page, Word files with their pictures
    • A folder as one .zip, after checking it fits

    Open the explorer guide →

  • dynamodb.py

    Amazon DynamoDB

    Tables and the items in them.

    • Every table's key, size, billing and cost
    • Scan, query and get items as plain tables
    • Which attributes the items hold, and their types
    • The read units each report used; scans stop early

    Open the DynamoDB guide →

  • bedrock_kb.py

    Amazon Bedrock Knowledge Bases

    Knowledge bases, what they retrieve, and the answers built on them.

    • An explorer window: every file, and how it was indexed
    • Settings in plain English, sync health and failed files
    • Search with highlighted passages; answers with each claim linked to its source
    • Compare search settings and measure retrieval hit rate

    Open the Knowledge Bases guide →

  • bedrock_chat.py

    Bedrock knowledge base chat

    A chat window on a knowledge base, with every setting in reach.

    • Pick the knowledge base and the model; answers stream in
    • Each answer's citations, sources, request and response, or only the search behind it
    • Any RetrieveAndGenerate setting, and the setup as Python, JSON or an AWS CLI command
    • Test a list of questions, and see which did better after a change

    Open the chat guide →

  • sagemaker_env.py

    Amazon SageMaker

    The notebook you're working in, and everything else SageMaker bills you for.

    • This notebook's type, cost so far and idle shutdown
    • Its CPU, memory, disk and GPU use right now
    • What fills the disk, and what's safe to clear
    • What's running in the region, and what looks forgotten

    Open the SageMaker guide →

  • opensearch.py

    Amazon OpenSearch

    Vector indexes in OpenSearch Service domains and Serverless collections.

    • Every vector field in plain English: size, engine, similarity
    • Whether the graphs fit in the memory the nodes have
    • Documents without a vector; zero or repeated vectors
    • The nearest neighbours of a question, a vector or a document

    Open the OpenSearch guide →

  • lambda_functions.py

    AWS Lambda

    Lambda functions, in one region or all of them.

    • An explorer window: every function, and its logs run by run
    • Every function's runtime, triggers, calls, errors and cost
    • Runtimes losing support, and functions anyone can call
    • Errors grouped by cause, from the function's own logs
    • Memory used and cold starts, and the size that would do

    Open the Lambda guide →

Every service works the same way

%pip install aws-analyzer gives you every service, and only needs boto3, which SageMaker already has. Each guide shows how to install it, from a bucket too when the notebook has no internet access.

from aws_analyzer import S3View  # or DynamoDBView, BedrockKBView, SageMakerView, OpenSearchView, LambdaView

ui = S3View()              # uses the notebook's IAM role; nothing to configure
ui.help()                  # every command, grouped by task; ui.help("name") shows one in full
s3 = ui.core               # the analyzer behind the view: returns data instead of a report

Every report reads the same way: cards with the numbers that matter (amber or red when something needs a look), findings that each end in what to do, the tables of detail, and a Next row of the commands worth running next, with the arguments already filled in. One click on any command in a report selects all of it, ready to copy.