DATA ENGINEERING / Amazon Web Services
AWS Spark Troubleshooting
Analyzes failed Spark workloads from supported AWS platforms through a managed regional MCP endpoint.
“A fully managed remote MCP server that provides specialized tools for troubleshooting Apache Spark applications on Amazon EMR, AWS Glue, and Amazon SageMaker Notebooks.”
01 / THE REASONING
Why this made the selection.
- Publisher-managed diagnostic service with a documented SigV4 proxy configuration.
- Documents failed-workload, region and log prerequisites rather than promising general Spark execution.
02 / THE REVIEW RECORD
What we actually inspected.
Source review has boundaries.
A clear record is more useful than a “safe” badge.
Material inspected
- https://github.com/awslabs/mcp/blob/8840003820a78dd018561d55c7ab12fcdd708464/src/sagemaker-unified-studio-spark-troubleshooting-mcp-server/README.md
- https://docs.aws.amazon.com/emr/latest/ReleaseGuide/spark-troubleshooting-agent-setup.html
Our findings
- Pinned README provides the exact proxy endpoint/service/profile arguments; current official setup adds cross-region inference disclosure. Managed backend code and inference retention were not independently inspected.
Not established by this review
- Server startup, authentication, live tool execution and security isolation were not tested.
- Published package/container parity with a source revision and hosted implementation internals were not tested.
The review applies to the material and revision named here. A newer upstream release can change its behavior.
03 / PUT IT TO WORK
Connect a server deliberately.
- Install Python3.10+, uv and AWS CLI; review the official setup template and create an appropriately scoped assume-role profile.
- Match endpoint, --region and profile to the workload region. The example uses us-east-1 and the documented 180-second proxy read timeout.
- Select a failed job and inspect diagnostic evidence; recommendations remain proposals until separately reviewed and tested.
Before you start
- Python3.10+, uv, AWS CLI and an authorized role/profile for the managed MCP service.
- A supported failed Spark workload with accessible event/history logs in the selected region.
Documented configuration example
Use the client and authentication context described in the setup guidance. Configuration formats vary between hosts.
{
"mcpServers": {
"aws-spark-troubleshooting": {
"command": "uvx",
"args": [
"mcp-proxy-for-aws@latest",
"https://sagemaker-unified-studio-mcp.us-east-1.api.aws/spark-troubleshooting/mcp",
"--service",
"sagemaker-unified-studio-mcp",
"--profile",
"smus-mcp-profile",
"--region",
"us-east-1",
"--read-timeout",
"180"
]
}
}
} 04 / FOLLOW THE EVIDENCE
The source trail.
Our notes are separate from the original resource.
Check upstream before adopting a new version.
https://github.com/awslabs/mcp/blob/8840003820a78dd018561d55c7ab12fcdd708464/src/sagemaker-unified-studio-spark-troubleshooting-mcp-server/README.md
Supports: summary, install, configExample, access, limitations, bestFor, review, compatibility, transports
https://docs.aws.amazon.com/emr/latest/ReleaseGuide/spark-troubleshooting-agent-setup.html
Supports: summary, install, configExample, access, limitations, bestFor, review, compatibility, transports