Connect Kubernetes cluster

Learn how to connect your Kubernetes cluster with Sedai's autonomous cloud management platform. Achieve continuous optimization and automation for your workloads.

Sedai integrates to individual Kubernetes clusters. To integrate a cluster, Sedai requires access to control plane APIs so it can view the current state of workloads and optionally make changes (such as modifying configurations like CPU, memory, etc). Sedai additionally needs access to monitoring data APIs to query metrics (such as usage, traffic, and performance) for its analysis and ML models.

Once connected, Sedai discovers stateless workloads, Deployments, StatefulSets, ReplicaSets, Jobs, CronJobs, Services, and Volumes.

Sedai currently supports Copilot and Autopilot modes for stateless workloads and partially supports stateful workloads; Datapilot mode is supported for all other workload types.

How to integrate

For clusters managed through a cloud provider, we recommend deploying Sedai's Smart Agent to ensure secure, uninterrupted access:

Sedai Smart Agent

If your organization has rules that limit egress access from within your cluster, Sedai requires explicit concessions to allow backend service from the Smart Agent.

Work with your security team to scan and validate the Smart Agent's container images to ensure they meet security needs.

Sedai also supports integrating clusters without an agent for AWS, Microsoft Azure, and Google Cloud Platform:

AKS Agentless SetupEKS Agentless SetupGKE Agentless SetupNode Optimisation Setup for AKS

If you run on-prem/self-managed clusters, you can also connect clusters deployed in Cloud Service Provider's environments.

Supported Providers

• AWS Elastic Kubernetes Service (EKS)

• Alibaba Container Service for Kubernetes

• Azure Kubernetes Service (AKS)

• DigitalOcean

• Google Kubernetes Engine (GKE)

• IBM Cloud Kubernetes Service

• OpenShift

• Oracle Cloud Infrastructure Container Engine for Kubernetes (OKE)

• Platform9

• Rancher

• VMware Tanzu