Katana

Inspiration

One of our teammates had prior exposure to the underlying ideas and features exposed by a platform like kubernetes and after a brief explanation, the others were eager to learn.

What it does

Katana uses an AI agent deployed alongside each of your kubernetes cluster which allows it to directly pull logs and performance metrics from each of your running pods, allowing it to remediate highly stressed servers while gaining an insight of the root of the problem through analysing logs.

How we built it

Using a tech stack that was new to everyone on the team, we employed django for the front end/backend and used minikube, kubectl, python, prometheus, django, docker and flask to ensure that we wrote clean, easily scalable code.

Challenges we ran into

The setup of the kubernetes cluster, and especially establishing the correct networking rules, event triggers and performance logging tools ate up a lot of our time as it was new to many of us. It is no joke when people say scalable code comes from a painful process of following best practices to a T.

Accomplishments that we're proud of

The way we built our codebase, from using best practices throughout setting up the tech stack, especially in areas of the kubernetes cluster & the general speed of us building this entire product to a working standard was something we were very proud of. Three of us had never touched kubernetes before the hackathon day and it was incredible to learn so much about a critical piece of tech so quickly. We're also proud of our UI design, none of us coming from a design background, we think the dashboard looks quite snazzy.

What we learned

The importance and agony of setting up your development environment with care. Creating easy to use scripts for easy deployment and destruction of docker containers must've saved us hours over the course of the hackathon.

What's next for Katana

More exploration into the world of kubernetes!

Built With

• django
• docker
• flask
• kubectl
• kubernetes
• prometheus
• python