Kubernetes Integration

The Kubernetes integration provides deep observability into your Kubernetes clusters. AtlasAI discovers namespaces, deployments, pods, and services, monitors cluster health, and enables Kubernetes-native remediation actions from runbooks.

Prerequisites

• A Kubernetes cluster (1.24+)
• kubectl access with permissions to create namespaces, service accounts, and DaemonSets
• Network connectivity from the cluster to the Tenant Plane

Setup

Step 1: Install the AtlasAI Kubernetes operator.

The operator handles service discovery, event collection, and metric aggregation:

kubectl create namespace atlasai
 
helm repo add atlasai https://charts.atlastechlab.com
helm install atlasai-operator atlasai/operator \
  --namespace atlasai \
  --set tenantUrl=<TENANT_PLANE_URL> \
  --set apiKey=<INTEGRATION_API_KEY>

Step 2: Deploy Edge Agents as a DaemonSet.

For node-level metrics and log collection:

helm install atlasai-agent atlasai/edge-agent \
  --namespace atlasai \
  --set tenantUrl=<TENANT_PLANE_URL> \
  --set apiKey=<AGENT_API_KEY> \
  --set kubernetes.enabled=true

Step 3: Verify the installation.

kubectl get pods -n atlasai

All pods should show Running status. Verify in the AtlasAI UI under Settings → Integrations → Kubernetes — the cluster should appear as “Connected”.

What Gets Discovered

CMDB Mapping

Kubernetes resources are automatically mapped to the AtlasAI CMDB:

• Clusters → Infrastructure CI
• Namespaces → Logical grouping
• Deployments → Service CI
• Pods → Instance CI (linked to their parent deployment)
• Nodes → Host CI (linked to Edge Agents if installed)

Kubernetes-Native Actions

Runbooks can include Kubernetes-specific actions:

• Scale a deployment (kubectl scale)
• Restart a deployment (kubectl rollout restart)
• Drain a node (kubectl drain)
• View pod logs (kubectl logs)
• Execute commands in containers (kubectl exec)
• Apply manifests (kubectl apply)

These actions execute via the AtlasAI operator, respecting RBAC and the configured trust level.

Troubleshooting