nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Stability Remote MCP from visual code

Dirk Verstraete_3ij0pe

i have succesfully implemented the integration, however when triggering mcp server it frequently hangs or does 1 task quit fast and then it stops responding which makes me restart the remote mcp manually. i know this is still beta but are there workarounds for this?

Comments

Isabella Thomas

Thanks for sharing, Dirk. I’ve noticed similar hangs in the beta. One workaround is to monitor the MCP logs in real time from VS Code and trigger tasks sequentially rather than in batch—this can reduce hangs. Also, setting a short auto-restart script for the MCP server can help avoid manual restarts.

Dirk Verstraete_3ij0pe

thanks for your quick response, i appreciate it!!

tasks are triggered sequentially, that's what supprises me, that 1 tasks works and a few seconds later it does not respond anymore.

in the mean time i found out that i was using an initial version 0.0.1 which never got update, i now stepped up to 1.0.0 but having the same problems.

i will implement an auto restart script, let you know if it works!

Taliah15

Hey @Dirk Verstraete

Congrats on getting the integration running! Since MCP is still in beta, these hang-ups are common—the server might freeze or quit after one task due to network, config, or resource hiccups. For now, check your logs, keep an eye on server resource usage, and consider upping timeout settings or optimizing your config to minimize manual restarts. You're not alone—tweaks and monitoring usually smooth things out!

Dirk Verstraete_3ij0pe

thanks for the timeout tip!

i added this now to the mcp config in visual code:
{
"chat.mcp.requestTimeout": 30000, // 30 seconds in milliseconds
"chat.mcp.connectionTimeout": 10000 // 10 seconds
}

Dirk Verstraete_3ij0pe

these settings has apparantly no impact, any suggestions for better implementation?

Taliah15

Increase timeouts: request Timeout to 120,000 m/s, connection Timeout to 15,000 m/s. Add a simple health check endpoint that returns server status and memory usage. Monitor server resources (CPU, memory, network) to avoid overload. Implement automatic health pings to detect hangs and auto-restart the MCP. Use community tools/extensions for stuck MCP refresh. Use community tools/extensions for stuck MCP refresh. Try this if it works kindly plz let me know. @Dirk Verstraete

Taliah15

@Dirk Verstraete Has your issue resolved or not? If your issue has resolved plz mark accepted answer. It's just a request.

Dirk Verstraete_3ij0pe

it is partially, writing comments is still an issue with mcp, but i use api for the comment writing.

Taliah15

Ohh! That's nice. But are you going any answer as marked answer or not? And what is your next step regarding the issue of comment writing with mcp? @Dirk Verstraete

Welcome!

It looks like you're new here. Sign in or register to get started.