Sentinal

Hey folks,

We are a pair of engineers who have been developing on Kubernetes for many places and different scales. We keep running into the same pain: logs scattered everywhere, unclear root causes, and everything always boils down to reading a wall of (poorly formatted) logging.

So we built Sentinal — a developer-focused tool that deploys AI agents *inside* your Kubernetes cluster to help you find and fix issues faster. It is designed to break issues down, explore them, with the guidance of human approved tools to find the root cause of issues.

Link to our website Sentinal

Unlike other AI tools in this space, we want developers to explore their issues and understand what is happening. We built Sentinal to look beyond logging and alerts but also navigate how your service interact with each other.

We’re building these agents to be like having a staff engineer who lives in your cluster 24/7. You can:

- Ask them questions about your system (they know your code and infrastructure).
- Port-forward to a live agent and interact with it via a Web UI.
- Watch what they’re doing — see the commands they run, the reasoning behind decisions, and what files they access (it’s all transparent).
- Ask the agent to help you test your integrations and suggests best approaches (provides helpful suggestions specific to your setup)

We’re focused on making this feel natural for developers:

- Helm-based install
- Safe by design
- RBAC roles are enabled by default, your pods can only run read commands and do not have the ability to create deployments
- Each Tool has to be approved by the user, you are in control of what happens
- It doesn’t just summarise logs — it digs into your business logic, and we aim to develop agents that can uncover hard to spot errors that similar tools would fail to spot. It uses your Github repositories that it finds relevant and tries to understand how your services work, and how they are deployed
- Think of tools that just feed off of prometheus and kubernetes events, they would only capture so much!

Right now we’re in open beta with a very simple agent - and its free!

As we start a funding round and get our project off of the ground, your feedback, good or bad, would be really helpful! And if you are interested in the project, and want to get involved in any way, get in contact!

Also, any feedback is welcome as we want to fine-tune our promotional materials, documentation, developer experience etc. in order to improve the experience for end users.

Thank you,

The Sentinal Team