A machine learning based recommendaiton engine using Prometheus metrics data, to optimize cost, performance, and compute capacity of Kubernetes workloads and clusters.

We built this working with our team of engineers, designers, and advisors. We are focused on cloud native workloads orchestrated by Kubernetes.

Most cloud billing optimization solutions look at your current cloud provider bill and shows you where your dollars go - within that cloud provider. What they don’t tell you is how that relates to your workload’s performance, distribution, or optimal compute capacity. Knowing how much you spend on cloud services won’t help you figure out whether that’s justified or not. Instead of looking at just one dimension, cost, we look at 3 dimensions - cost, performance, and capacity. This is a hard problem for humans to solve. It would take 2-3 humans to continually look at these dimensions to figure out whether they use spot instances, let alone when, how many, etc. When each hour/day the deployment frequency changes, the workload behavior changes and therefore the cost/performance ratio changes. The customers that spend any where between 4 figure (1000s of $) to 6 figures (e.g $100K or more) have no idea how to optimize their spend without affecting the SLA's of their workloads. These are the types of questions customers are asking us: https://www.katen.ai/quotes

Thank you for all the upvotes. It’s great to see people are interested in this problem. We are also heartened that several people signed up for Early Access. Your support and encouragement means a lot to us at this trying time. Be safe and be well!

Katen

Machine learning based Kubernetes optimization engine

Machine learning based Kubernetes optimization engine

Katen

Machine learning based Kubernetes optimization engine

Machine learning based Kubernetes optimization engine

Do you use Katen?

Do you use Katen?

You might also like