Architecture

How Ankra Works

A lightweight agent connects your cluster to the Ankra platform. No firewall changes, no privileged access, no vendor lock-in.

Simple, Secure Architecture

Your cluster stays in your environment. The agent handles communication.

Your Cluster

EKS, GKE, AKS, on-prem, local

Ankra Agent
Lightweight, outbound-only

Ankra Platform

UI, AI, GitOps engine

Dashboard & Stack Builder
AI Assistant
Git Sync
Your Git Repository
Source of truth for configs

Outbound-Only Agent

The agent connects to Ankra's platform - your cluster never exposes any ports. No firewall changes, no ingress configuration.

Your Data, Your Control

Sensitive data like secrets are never sent to the AI. Configurations live in your Git repo. You can remove the agent anytime.

Works Everywhere

Cloud providers (EKS, GKE, AKS), on-premises clusters, local development (Minikube, Kind). If it runs Kubernetes 1.24+, it works.

Get Started in 4 Steps

From zero to managing your cluster in under 10 minutes.

01

Import Your Cluster

Run a single Helm command to install the lightweight Ankra agent. Works with any Kubernetes cluster - EKS, GKE, AKS, on-prem, or local.

  • Agent connects outbound only - no firewall changes
  • No privileged access required
  • Takes less than 5 minutes
helm install ankra-agent \
  oci://registry.ankra.cloud/ankra/ankra-agent \
  --set config.token=YOUR_TOKEN
02

Connect Your Git Repository

Link your GitHub repository to enable GitOps. All configurations are stored in your repo as standard Helm values and Kubernetes manifests.

  • Standard Helm charts - no proprietary formats
  • Your Git repo is the source of truth
  • Full audit trail of all changes
# Your repo becomes the source of truth
my-repo/
├── manifests/
│   ├── namespace.yaml
│   └── rbac.yaml
├── addons/
│   ├── prometheus/values.yaml
│   └── grafana/values.yaml
└── import-cluster.yaml
03

Build & Deploy Stacks

Use the visual Stack Builder to compose your infrastructure. Define dependencies, configure values, and deploy with one click.

  • Drag-and-drop Helm charts and manifests
  • Visual dependency management
  • Deploy via UI, CLI, or Git push
# Or use the CLI
ankra apply stack monitoring

# Or push to Git
git push origin main  # Auto-deploys
04

Monitor & Debug with AI

Get real-time visibility into your clusters. When something breaks, ask the AI - it already has the logs, events, and manifests.

  • AI sees logs, events, manifests together
  • Automatic incident analysis on alerts
  • Context-aware troubleshooting
You: "Why is the payment-service crashing?"

AI: "The pod is OOMKilled. Memory limit 
is 256Mi but peak usage hit 312Mi. 
The 'backend' Stack deployment 2h ago 
increased the connection pool from 
10 to 50, causing higher memory usage."

What Data Flows Where

Transparency about what the agent sends and what stays in your cluster.

Sent to Ankra Platform

  • Resource metadata (names, namespaces, labels)
  • Pod/deployment status and events
  • Container logs (for AI analysis)
  • Helm release information
  • Resource health and metrics

Never Sent

  • Secret values or credentials
  • ConfigMap sensitive data
  • Service account tokens
  • Your application data
  • Anything marked sensitive

Ready to import your first cluster?

Free tier available. No credit card required.

Get Started FreeQuickstart Guide