I Built a Local-First Kafka GUI From Scratch to Handle Massive Clusters

Kezlo Smith — Sat, 20 Jun 2026 04:03:15 +0000

Every data engineer or backend developer who has worked with Apache Kafka eventually runs into the same wall: tooling.

For a long time, I felt stuck between two extremes: older, native desktop apps that look like they were built two decades ago, or beautiful modern web consoles that lock up or timeout the second a cluster scales to hundreds of topics and gigabytes of metadata.

I hit that breaking point on my own projects. I was tired of synchronous network bottlenecks freezing my interface during critical debugging sessions, so I decided to build the tool I actually wanted to use: Watermark.

Why I Built Watermark

Watermark is a 100% local-first, open-source desktop application designed to handle complex, heavy Kafka environments smoothly. I built it from the ground up to be lightweight, fast, and entirely local.

The Tech Stack

I went with a split architecture to maximize efficiency:

Go on the backend to handle high-velocity cluster communication, metadata parsing, and authentication heavy lifting.
React on the front end to craft a clean, high-observability interface inspired by modern IDE design patterns.

Because it runs entirely locally, your cluster credentials, SASL, SSL, or AWS IAM setups never leave your machine.

Features I Focused On

1. Consumer Lag Ranking (No More Hunting)

Instead of forcing you to click into individual consumer groups to find out why a pipeline is backed up, I built a dedicated monitoring dashboard. It surfaces a global view of topic and group lag instantly, ranking them by priority.

2. Context-Aware Topic Annotations

In large environments, figuring out who owns a topic is a daily nightmare. Watermark introduces native metadata tagging. You can document producers, consumers, and operational context directly on the topic view.

3. Zero-Freeze Multi-Cluster Management

You can jump between development, staging, and read-only production environments seamlessly. I optimized the metadata lifecycle so that massive topic counts won’t time out your workspace.

Current Limitations

Right now, the pre-compiled release is macOS only (Universal DMG). However, because the underlying Go/React codebase is cross-platform, Windows and Linux support is entirely possible.

If you are on Linux or Windows, you can compile it from source, or check out the active tracking issue on my GitHub to vote on native packaging priorities!

Check it out 🚀

The project is entirely open-source under the Apache 2.0 license.

Landing Page: watermark.kezlo.dev
GitHub Repository: Kezlo98/watermark

I'd love to know: What does your current Kafka debugging workflow look like, and what is the biggest missing feature in your current tools? Let's chat in the comments!

DEV Community: Kezlo Smith