MWAA

You can run an MCP server to manage MWAA environments and interact with Airflow through a unified, programmable interface. This lets you automate environment operations, DAG management, and routine tasks from an AI-assisted workflow.

How to use

You will run one or more MCP client configurations that connect to your MWAA MCP server. Use these clients to manage environments, inspect DAGs and tasks, view connections and variables, and generate access tokens for secure CLI or web access. Start by selecting a local or containerized client configuration that points to your MWAA MCP server, then issue natural language commands that map to AWS and Airflow actions.

How to install

# Prerequisites
- AWS credentials configured with MWAA permissions
- Python 3.10 or higher
- uvx installed for package management (recommended)

# Basic setup steps (local development)
# 1) Ensure you have uvx available in your environment
# 2) Prepare a configuration file for your MCP client
# 3) Run the MCP client to connect to the MWAA MCP server
```} ,{

Configuration and usage notes

The MWAA MCP server supports local and remote client configurations. Use a stdio-based setup to run the server locally via a command like uvx path/to/mwaa-mcp-server, or run the server inside a container using Docker. Both configurations should expose the AWS region and profile you want to operate under, and you can enable read-only mode for safer operations.

Configuration and usage notes

The MWAA MCP server supports local and remote client configurations. Use a stdio-based setup to run the server locally via a command like uvx path/to/mwaa-mcp-server, or run the server inside a container using Docker. Both configurations should expose the AWS region and profile you want to operate under, and you can enable read-only mode for safer operations.

Troubleshooting and best practices

If you encounter authentication errors, verify your AWS credentials and region, and ensure the MWAA permissions are in place. For connection timeouts, check network reachability to MWAA endpoints and confirm the environment is in AVAILABLE state. For DAG import errors, review import logs and ensure dependencies are present.

Available tools

listEnvironments

List all MWAA environments in the account with their configurations.

describeEnvironment

Describe a specific MWAA environment by name to view details.

createEnvironment

Create a new MWAA environment with the specified configuration.

updateEnvironment

Update an existing MWAA environment to change its settings.

deleteEnvironment

Delete an MWAA environment to clean up resources.

generateCliToken

Generate a CLI access token to authenticate against MWAA CLI.

generateWebLoginToken

Generate a token for web UI login to MWAA.

listDags

List DAGs within an MWAA environment.

viewDag

Retrieve the source or definition of a DAG.

triggerDag

Trigger a DAG run with optional configuration.

monitorDagRuns

Monitor ongoing or recent DAG runs for status.

getDagLogs

Fetch logs for a specific DAG run or task instance.

listTaskInstances

Track status of individual task instances within a DAG run.

getConnections

View Airflow connections configured in the environment.

getVariables

View Airflow variables for a given environment.

getImportErrors

Diagnose DAG parsing/import errors.