Handling Long-Running Tasks in Python 3 MCP Server with SSE: Avoiding Client Timeouts and Session Closures

[fish2018]

Pre-submission Checklist

• I have checked that this question would not be more appropriate as an issue in a specific repository
• I have searched existing discussions and documentation for answers

Question Category

• Protocol Specification
• SDK Usage
• Server Implementation
• General Implementation
• Documentation
• Other

Your Question

I am currently developing an MCP (Model Context Protocol) server using Python 3 and have encountered a challenge. The platform exclusively supports SSE (Server-Sent Events), and the execution of certain tools can be quite time-consuming. If I opt for synchronous execution, there's a risk of client timeouts due to the extended processing time. On the other hand, if I choose asynchronous execution by initially sending a "processing" status to the client, there's a possibility that the client session may terminate before the tool completes its task. This is particularly problematic in scenarios such as deploying services to Kubernetes or automating the purchase of cloud resources, where operations can be lengthy.

I am seeking advice on how to effectively manage these long-running tasks while ensuring that the client remains informed or can retrieve the results once the task is finished. Any insights or best practices for handling such situations would be greatly appreciated, as I plan to share this query with a technical community for further assistance.