Debugging your MCP integration

As the list of available tools keeps growing, sooner or later something will not work and some debugging becomes necessary. In this post I look into ways to troubleshoot your MCP integration.

Let's dive in!

To understand what we need to debug you need to be aware of the architecture of an MCP integration. It follows a client-server architecture, where AI models can request data through a host with an MCP client (e.g. GitHub Copilot, Claude Desktop, …) from MCP servers, which then retrieve relevant information from local or remote sources.

This means that when a problem occurs that there (at least) 2 places to look at.

Debugging the MCP client

Let’s start by looking into the client. The way you need to debug the client is completely dependent on the host. I’ll focus on GitHub Copilot and Claude Desktop.

GitHub Copilot

When VSCode encounters an error while trying to interact with an MCP server, you get a red error indicator in the Chat window:

Click on the icon and choose Show Output from the dropdown menu.

This will open up the logs in the Output window.

Remark: Another option is to run MCP: List Servers from the Command Palette, select the server, and then choose Show Output.

Claude Desktop

The ‘go to’ location in case of troubles with MCP support in Claude Desktop is the MCP log file. This file can be found by going to File –> Developer –> Open MCP Log file… in Claude Desktop:

This log file contains a lot of information:

Server connection events
Configuration issues
Runtime errors
Message exchanges

Debugging the MCP server

The way to troubleshoot and debug a specific MCP server will completely depend on the MCP server implementation. So there are no strict rules or guidelines I could share here.

But I do want to point out the MCP inspector tool. This tool allows you to test and debug your MCP server.

When using SSE

When using SSE, you can install and run the tool directly through npx:

npx @modelcontextprotocol/inspector

This will start the tool on a specific local URL ( http://127.0.0.1:6274 in our example).

Browse to that location to open the Inspector.

Now we can connect to our MCP server by specifying the SSE transport and specifying the URL of our MCP server:

Click on Connect afterwards connect to the MCP server.

Now you can go to Tools and click on List Tools to get the list of available tools and even give a tool a try:

When using STDIO

When using STDIO, you can include the command to start the MCP server when installing and running the tool:

npx @modelcontextprotocol/inspector <command> <arg1> <arg2>

The rest of the experience remains the same.

More information

Debugging - Model Context Protocol

Use MCP servers in VS Code (Preview)

Inspector - Model Context Protocol

Podman– Command execution failed with exit code 125

After updating WSL on one of the developer machines, Podman failed to work. When we took a look through Podman Desktop, we noticed that Podman had stopped running and returned the following error message: Error: Command execution failed with exit code 125 Here are the steps we tried to fix the issue: We started by running podman info to get some extra details on what could be wrong: >podman info OS: windows/amd64 provider: wsl version: 5.3.1 Cannot connect to Podman. Please verify your connection to the Linux system using `podman system connection list`, or try `podman machine init` and `podman machine start` to manage a new Linux VM Error: unable to connect to Podman socket: failed to connect: dial tcp 127.0.0.1:2655: connectex: No connection could be made because the target machine actively refused it. That makes sense as the podman VM was not running. Let’s check the VM: >podman machine list NAME ...

Cache stampede: when our cache turned against us

While investigating some performance issues, we ran into an ASP.NET Core API that cached a fairly expensive aggregation query for 60 seconds. Under normal load, that was fine: one request rebuilds the cache, everyone else reads from it. Under peak load, dozens of requests would arrive in that same expiry window, all see a cache miss, and all fire the same expensive query in parallel. The database didn't like that. That was the moment when our caching layer stopped helping and started hurting. A burst of requests comes in at the same time, all miss the cache, and all go hammer the database or the downstream API at once. That's a cache stampede . The cache was supposed to protect our backend, and for a few hundred milliseconds it did the opposite. Why this happens IMemoryCache.GetOrCreate (and its async sibling) looks like it protects you, but it doesn't add any locking on its own. Look at the naive version: public async Task<Report> GetReportAsync(string key) ...

A complex system designed from scratch never works

A few years ago, I worked as an architect on a big mainframe rewrite. I still count it as one of my failures. Not because the technology was wrong, but because I couldn't convince the management team to simplify the approach. Years later, the organization is still struggling to get the new system up and running. I left the project at the time, because I couldn't put my name behind an approach that would take very long and cost a lot of money without a working system to show for it along the way. Gall’s Law That memory keeps coming back to me, because it's a textbook case of Gall's Law playing out in real life. Gall's Law , from John Gall's Systemantics , states it plainly: A complex system that works is invariably found to have evolved from a simple system that worked. A complex system designed from scratch never works, and it cannot be patched to make it work. You have to start over with a simple system that works. What does that mean in practice,...

The art of simplicity

Search This Blog