Zesty Kompass autonomously rightsizes K8s pods, persistent volumes, and spot migration in under 40 seconds.
Kompass continuously rightsizes pod CPU and memory based on real-time usage, autoscales persistent volumes (rare among rightsizing tools), and migrates spot instances with sub-40-second replacement. ML models drive ongoing adjustments without code changes, and the Commitment Manager runs micro-commitment strategies for AWS Savings Plans and RIs to track usage shifts more tightly than annual commits.
Who it's for. Platform teams of 5 to 20 engineers running Kubernetes on AWS who want autonomous cost optimization without changing application code. Scenario: a cluster has 200 pods at 12 percent average CPU utilization, Kompass adjusts requests and limits to match actual usage, and cluster spend drops 35 percent in two weeks while underutilized EBS volumes shrink without downtime.
Tradeoffs. AWS and Kubernetes only; no GCP or Azure. Commitment Manager is AWS Savings Plans and RIs only. Autonomous patching of live workloads needs careful guardrails. Pricing is opaque: Kompass is monthly with unpublished prices, Enterprise is sales-only. Acquired by DoiT International, so the roadmap now ties to DoiT's strategy.
Compare: CAST AI, Spot by NetApp, StormForge, Karpenter
