---
title: "Discussion on \"MCP in Production, Part 6: Java vs Go, Measured\" | Hashnode"
description: "Discussion on \"MCP in Production, Part 6: Java vs Go, Measured\". TL;DR\n\nSame hexagon, same tools, same k6 scenario, two runtimes. Both servers handled the agent burst with zero failures and stayed far below the SLO. Go answered with a ~5× lower p95 (2.6 ms vs 13.8"
url: "https://hashnode.com/posts/mcp-in-production-part-6-java-vs-go-measured/6af883f3c928a577f810f3c6"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [MCP in Production, Part 6: Java vs Go, Measured](https://cbuelvasc.hashnode.dev/mcp-in-production-part-6-java-vs-go-measured?utm_source=hashnode&utm_medium=feed)

TL;DR Same hexagon, same tools, same k6 scenario, two runtimes. Both servers handled the agent burst with zero failures and stayed far below the SLO. Go answered with a ~5× lower p95 (2.6 ms vs 13.8

9h ago · 6 min read · cbuelvasc.hashnode.dev  
Read the article

-   [#java](https://hashnode.com/tag/java)
-   [#golang](https://hashnode.com/tag/golang)
-   [#performance](https://hashnode.com/tag/performance)
-   [#mcp](https://hashnode.com/tag/mcp)

## Written by

[Carmelo Buelvas Comas](https://hashnode.com/@cbuelvasc)

@cbuelvasc · Colombia

  
[View profile](https://hashnode.com/@cbuelvasc)

Senior Backend Engineer · Java, Go & AI (MCP) · Building production-grade systems

Senior Software Engineer with 14+ years building backend systems in Java (8 to 25) and Spring Boot, now also working with Go. I write about MCP, AI agents, cloud-native architecture, observability and taking systems to production. Based in Colombia.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.