AWS Documentation MCP Server
MCP setupDevOps and cloud
FreeMCP setupDevOps and cloud
Query Honeycomb traces, datasets and SLOs from your agent through Honeycomb's hosted MCP server, signed in with OAuth.
npx workspacetemplates add honeycomb-mcp --tool claude-codeWrites 1 file to your project root
{
"mcpServers": {
"honeycomb": {
"type": "http",
"url": "https://mcp.honeycomb.io/mcp"
}
}
}npx workspacetemplates add honeycomb-mcp --tool codexWrites 1 file to ~/.codex/
[mcp_servers.honeycomb]
url = "https://mcp.honeycomb.io/mcp"npx workspacetemplates add honeycomb-mcp --tool cursorWrites 1 file to .cursor/
{
"mcpServers": {
"honeycomb": {
"url": "https://mcp.honeycomb.io/mcp"
}
}
}npx workspacetemplates add honeycomb-mcp --tool gemini-cliWrites 1 file to .gemini/
{
"mcpServers": {
"honeycomb": {
"httpUrl": "https://mcp.honeycomb.io/mcp"
}
}
}npx workspacetemplates add honeycomb-mcp --tool copilotWrites 1 file to .vscode/
{
"servers": {
"honeycomb": {
"type": "http",
"url": "https://mcp.honeycomb.io/mcp"
}
}
}Honeycomb runs the server itself, which means you install nothing locally. Once connected, an agent can run queries on your event data, fetch a trace, look up column definitions and review SLOs while it works on a slow endpoint or an error spike.
You sign in with OAuth in the browser. The consent screen lets you choose read-only access, so the agent can investigate without creating boards or triggers. Teams in the EU region use the eu1 address instead of the one in this setup.
you › In the checkout dataset, which route had the biggest rise in p99 duration since 13:00, and show one slow trace.
agent › POST /checkout/confirm: p99 went from 0.9 s to 3.4 s. Slow trace 7f3a...: 2.8 s of the 3.1 s sits in the inventory.reserve span, waiting on a database call.MCP setupDevOps and cloud
FreeMCP setupDevOps and cloud
FreeMCP setupDevOps and cloud
FreeMCP setupDevOps and cloud
Free