Excessive Startup Requests Causing Early Shutdowns?

[resolutionathens]

Hi team,

I've been experiencing mcp-atlassian shutting down prematurely in Claude Desktop. Reviewing the logs, it appears the integration performs a significant amount of activity on startup, even before any explicit user interaction. It is grabbing Jira issues and Confluence spaces without being asked.

I suspect this initial burst of activity might be leading to too many requests or another internal issue causing the early shutdown and preventing the use of Atlassian integrations.

Has anyone else encountered this, or could the team shed light on the startup behavior of mcp-atlassian? I'm wondering if a more demand-driven initialization could address this.

Thanks!

[sooperset]

Hi @resolutionathens,

Thank you for reporting this issue. You've identified a key problem I've been working on - the integration was performing too many preemptive API calls during initialization. I've addressed this in the recent v0.8.0 release by refactoring the server code with FastMCP, which drops these resource requests due to its limitations.

If you're still experiencing issues after updating, please let me know with your environment details!