Skip to content

Oracle Cloud Free Tier for OKE Kubernetes

This cluster maximizes Oracle Cloud Infrastructure’s Free tier resources.

pie showData
    title OCI Free Resource Allocation
    "Worker 1 (2 OCPU, 12GB)" : 50
    "Worker 2 (2 OCPU, 12GB)" : 50

OCI Always Free provides 4 OCPUs and 24GB RAM for Ampere A1 instances. This cluster divides these resources across two worker nodes, while the Control Plane is managed by OCI (Free for Basic Cluster).

NodeOCPUsRAMPurpose
Control Plane--OKE Basic Cluster (Managed)
Worker 1212GBWorkloads
Worker 2212GBWorkloads

Total: 4 OCPUs, 24GB RAM (exactly at the limit)

graph LR
    subgraph Free["Free Tier (4 OCPU, 24GB)"]
        W1[Worker 1<br/>2 OCPU<br/>12GB]
        W2[Worker 2<br/>2 OCPU<br/>12GB]
    end

    subgraph Managed["OCI Managed"]
        CP[Control Plane<br/>Basic Cluster]
    end

    subgraph Used["Used: 4 OCPU, 24GB"]
        Total[100% Utilized]
    end

    W1 --> Total
    W2 --> Total

Ampere A1 Flex instances are ARM64-based. Container images must support the linux/arm64 architecture.

  • 1 VCN with public and private subnets
  • 1 Internet Gateway
  • 1 NAT Gateway (for private subnet outbound access)
  • 1 Network Load Balancer (Always Free includes 1 flexible NLB)
  • 200GB total block volume storage
  • Boot volumes count against this limit
  • Each node uses a 50GB boot volume (100GB total used)
  • Remaining storage available for Persistent Volumes via OCI CSI
  • 10TB outbound data transfer per month
  • Unlimited inbound

This cluster maximizes free resources, but note that NAT Gateway incurs a small cost if used heavily (though usually negligible).

OCI Always Free includes 1 Flexible Network Load Balancer. This cluster uses it to provide a stable public IP for ingress traffic:

flowchart LR
    Internet((Internet)) -->|:443| NLB[Network Load Balancer]
    NLB -->|TCP forward| Ingress[Ingress Node<br/>hostPort]
    Ingress --> Envoy[Envoy Pod]
    Envoy --> Apps[Applications]

Benefits of using the NLB:

  • Stable public IP (doesn’t change if ingress node is replaced)
  • Health checks on backend
  • Can add more ingress nodes later for HA

The OCI CSI driver allows dynamically provisioning Block Volumes for persistent storage, using the remaining free tier storage capacity.

Terraform enforces the correct instance shapes. Do not manually resize instances in the OCI Console.

Ampere A1 capacity varies by region. US-Ashburn-1 and EU-Frankfurt-1 typically have better availability. If provisioning fails with “Out of Capacity,” try a different availability domain or region.

Check your tenancy limits in the OCI Console under Governance > Limits, Quotas and Usage. Filter by “compute” to see Ampere A1 availability.