OCI Flexible Shapes Terraform – Boost Performance with Custom CPU/Memory

When building cloud-native applications or migrating existing workloads to Oracle Cloud Infrastructure (OCI), flexibility is key. That’s where OCI flexible shapes come into play — giving you complete control over the number of OCPUs and memory allocated to your virtual machines. In this guide, you’ll discover how to leverage OCI flexible shapes with Terraform to achieve an ideal balance between performance and cost.

OCI flexible shapes allow you to launch compute instances with as few as 1 OCPU and 1 GB of memory and then dynamically adjust the configuration as your workload evolves. Unlike fixed shapes (e.g., VM.Standard2.1, VM.Standard3.1), OCI flexible shapes like VM.Standard.E4.Flex allow you to define the number of OCPUs and memory size, giving you greater control over resource allocation and cost.

In this hands-on tutorial, you’ll see how to:

Deploy an OCI compute instance using flexible shapes
Configure the number of OCPUs and memory with Terraform variables
Save on cloud costs by provisioning only what your workload needs
Adjust resources over time without disrupting the workload

Why Use OCI Flexible Shapes?

Here are a few reasons why flexible shapes are a game-changer in OCI:

✅ Optimize cloud spending – only pay for the CPU and RAM you use
✅ Improve resource efficiency – adjust instance configuration on the fly
✅ Deploy faster – provision right-sized infrastructure in seconds
✅ Eliminate waste – ideal for test environments, CI/CD runners, and variable workloads

In the video tutorial (below), I demonstrate how to define and deploy an instance using Terraform with OCI flexible shape. You’ll see how to configure OCPUs and memory directly in your .tf files and how to use variables to simplify infrastructure changes.

Practical Benefits of Flexible Shapes in OCI

Using OCI flexible shapes like VM.Standard.E4.Flex allows you to allocate exactly the amount of OCPUs and memory your workload needs. This results in better cost optimization and resource efficiency compared to fixed-shape instances. In production environments, this flexibility helps fine-tune performance while avoiding overprovisioning.

For example, if your application only needs 2 OCPUs and 8GB of RAM, there’s no need to pay for unused capacity. With Terraform, it’s easy to make such granular configurations repeatable and version-controlled across environments.

Sample Terraform Configuration

Here’s a sample Terraform snippet that provisions an OCI compute instance with a flexible shape:


resource "oci_core_instance" "foggykitchen_flex_vm" {
  (...)
  shape = "VM.Standard.E4.Flex"

  shape_config {
    ocpus         = var.ocpus
    memory_in_gbs = var.memory
  }
  (...)
}

You can adjust the ocpus and memory_in_gbs values based on the current needs of your application — making it ideal for dynamic workloads or environments with unpredictable load patterns.

💡 This screenshot comes from OCI Cloud Shell
The image shows how Terraform code is executed directly in the OCI Cloud Shell environment. It’s a convenient and secure way to deploy flexible compute shapes (e.g. VM.Standard.E4.Flex) in Oracle Cloud without setting up your local environment. The code configures OCPUs and memory dynamically — perfect for DevOps and self-study labs.

Terraform Tip: Using dynamic_block for Flexible Shapes

When provisioning OCI compute instances with OCI flexible shapes (like VM.Standard.E4.Flex), you can use Terraform’s dynamic block to configure shape_config only when needed. Here’s a practical example:


resource "oci_core_instance" "foggykitchen_flex_vm" {
(...)
  shape               = var.shape
  
  dynamic "shape_config" {
    for_each = local.is_flexible_shape ? [1] : []
    content {
      ocpus         = var.ocpus
      memory_in_gbs = var.memory
    }
  }    
(...)    
}

# Dictionary Locals
locals {
  compute_flexible_shapes = [
    "VM.Standard.E3.Flex",
    "VM.Standard.E4.Flex",
    "VM.Standard.A1.Flex",
    "VM.Optimized3.Flex"
  ]
  is_flexible_shape = contains(local.compute_flexible_shapes, var.shape)
}
(...)

This dynamic approach ensures your code remains DRY and modular – shape configuration is only applied for flexible shapes like VM.Standard.E4.Flex.

E3 is Out, E4 is In

The VM.Standard.E3.Flex shape is being deprecated, and Oracle Cloud now recommends VM.Standard.E4.Flex for general-purpose workloads. E4 brings better price-performance ratio and is based on newer AMD EPYC processors.

If you’re still using OCI Flexible Shape E3 in your Terraform templates, consider updating your defaults to:


variable "Shape" {
  default = "VM.Standard.E4.Flex"
}

OCI Flexible Shapes with Terraform – What’s Next?

If you’re interested in mastering other Terraform automation topics in OCI, check out these courses:

You can also explore the source code examples and exercises in this GitHub repository (if applicable, based on your current setup).

If you want to go beyond compute flexibility, check out my full OCI with Terraform course

🚀 Go Beyond Flexible Shapes: Master Full OCI Automation

This blog gave you a glimpse into OCI compute flexibility.
In the full flagship Terraform course, you’ll:

- Automate complete OCI networks, compute, and load balancers
- Deploy production-ready environments step by step
- Gain lifetime access to 10+ real-world labs

🔒 Lifetime • ⏱️ Self-paced • 🧪 Real labs

Why Use OCI Flexible Shapes?

Practical Benefits of Flexible Shapes in OCI

Sample Terraform Configuration

Terraform Tip: Using dynamic_block for Flexible Shapes

E3 is Out, E4 is In

Read more about OCI & Terraform

OCI Flexible Shapes with Terraform – What’s Next?

🚀 Go Beyond Flexible Shapes: Master Full OCI Automation

Check also other courses:

1 Comment

Leave A Reply Cancel reply

🚀 Go Beyond Flexible Shapes: Master Full OCI Automation

Login with your site account

Register a new account