AWS Glue Interactive Sessions Kernel

1.0.9 verified Fri May 01 auth: no python

Provides a Jupyter kernel for AWS Glue Interactive Sessions, enabling interactive development of PySpark, Ray, and SQL jobs from any Jupyter environment (including IDEs). Current version 1.0.9. Released occasionally, no fixed cadence.

pip install 'aws-glue-sessions>=1.0.0' jupyter

Common errors

error KeyError: 'glue_pyspark' ↓

cause Kernel spec not installed; user tried to select a kernel that doesn't exist.

fix

Reinstall the package in the correct Python environment: pip install --user 'aws-glue-sessions>=1.0.0' and verify with jupyter kernelspec list.

error NoCredentialsError: Unable to locate credentials ↓

cause AWS credentials not configured. The kernel fails to start a Glue session.

fix

Set environment variables AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY (and optionally AWS_SESSION_TOKEN) or configure AWS CLI.

error ModuleNotFoundError: No module named 'aws_glue_sessions' ↓

cause User mistakenly tries to import the package as a Python module instead of using it as a Jupyter kernel.

fix

Do not import the package. Install it and launch Jupyter to select the kernel from the interface.

Warnings

gotcha The kernel spec is installed during pip install, not via jupyter kernelspec install. If you see 'No kernel named glue_pyspark', reinstall the package or check your Jupyter installation is on the same Python environment. ↓

fix Ensure you are installing in the same Python environment as Jupyter: run 'which jupyter' and 'pip list | grep aws-glue-sessions'. Reinstall if needed.

gotcha When using from an IDE (VS Code, PyCharm), the kernel spec must be registered globally. The package uses jupyter_client.install_kernel_spec() which may require write permissions to Jupyter's kernel directory. ↓

fix Install with --user flag: pip install --user 'aws-glue-sessions>=1.0.0' or run as admin. On some systems, use 'python -m jupyter_client.kernelspec --user' manually.

deprecated The old import path 'from aws_glue_sessions import ...' is no longer supported. The package does not expose Python modules; it only registers kernel specs. ↓

fix Do not import anything from the package; simply install and use the kernel from Jupyter.

gotcha Interactive Sessions require AWS credentials to be configured (via environment variables, AWS CLI, or IAM role). The kernel will fail to start with 'No credentials provided' if not set. ↓

fix Set AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_SESSION_TOKEN (if using temp creds) or configure ~/.aws/credentials before launching Jupyter.

Quickstart

Install the kernel and verify it appears in Jupyter. Then launch a notebook and select the Glue Interactive Sessions kernel.

pip install 'aws-glue-sessions>=1.0.0' jupyter
python -m sp流水线 install --user --symlink
jupyter kernelspec list  # should list glue_pyspark, glue_spark, glue_ray
jupyter notebook  # launch Jupyter, select Glue PySpark kernel