When should I choose Code Engine over IKS?

Code Engine is right when your workload is essentially stateless containers without significant per-pod customization (no DaemonSets, no node-level configuration, no PersistentVolume needs). IKS is right when you need Kubernetes-native primitives, complex multi-service routing, custom operators, or persistent storage at scale. For 'I have a web app and a worker', Code Engine is dramatically simpler.

How does Code Engine pricing work?

You pay per second of execution for applications and jobs, plus a small per-1000-request fee. There's a generous free tier each month. Idle applications scale to zero and stop billing. This means low-traffic apps cost almost nothing, while high-throughput apps cost more than a comparably-sized always-on Droplet would — model both before committing.

IBM Code Engine Config Builder

ComputeIBM Cloud

Build Code Engine project configurations with applications, jobs, secrets, and event subscriptions.

Last verified: May 2026

IBM Cloud Code Engine Configuration

Build Code Engine project configurations with applications, jobs, secrets, and event subscriptions for serverless workloads.

Required Fields

projectNameregion

{
  "projectName": "prod-microservices",
  "resourceGroup": "prod-resources",
  "region": "us-south",
  "applications": [
    {
      "name": "api-gateway",
      "imageReference": "us.icr.io/prod-ns/api-gateway:v2.1.0",
      "cpu": "1",
      "memory": "4G",
      "minScale": 2,
      "maxScale": 10,
      "concurrency": 100,
      "port": 8080,
      "envVariables": [
        { "name": "NODE_ENV", "value": "production" },
        { "name": "DB_URL", "valueFrom": { "secretKeyRef": { "name": "db-credentials", "key": "url" } } }
      ],
      "scalingConcurrencyTarget": 80,
      "requestTimeout": 300,
      "healthCheck": {
        "type": "http",
        "path": "/health",
        "initialDelay": 5
      },
      "visibility": "public"
    }
  ],
  "jobs": [
    {
      "name": "data-processor",
      "imageReference": "us.icr.io/prod-ns/data-processor:v1.5.0",
      "cpu": "2",
      "memory": "8G",
      "maxExecutionTime": 3600,
      "retries": 3,
      "arraySize": 10,
      "envVariables": [
        { "name": "BATCH_SIZE", "value": "1000" }
      ]
    }
  ],
  "secrets": [
    {
      "name": "db-credentials",
      "type": "generic",
      "data": {
        "url": "<database-url>",
        "password": "<database-password>"
      }
    }
  ],
  "eventSubscriptions": [
    {
      "name": "cos-trigger",
      "type": "cos",
      "source": {
        "bucket": "incoming-data",
        "prefix": "uploads/",
        "events": ["write"]
      },
      "destination": {
        "type": "job",
        "name": "data-processor"
      }
    }
  ]
}

Generated Output

Output will appear here...

About This Tool

IBM Cloud Code Engine is the platform's serverless compute offering — run containers, source code, or batch jobs without managing a Kubernetes cluster. The Code Engine Config Builder generates project, application, job, secret, and event subscription specs. Output is Terraform-ready and structured so a Code Engine project's components live in one declarative file.

Real-World Scenario

A nightly data-import process has been running on a t2.medium Droplet that's idle 23 hours a day — $20/month for an hour of actual work. You migrate it to a Code Engine job triggered nightly. The new cost is under $1/month and the Droplet gets decommissioned. The team adopts Code Engine for similar batch workloads, dropping a layer of infrastructure to maintain.

When to Use This Tool

•Defining a new Code Engine project with multiple applications (an HTTP API, a worker job) that share environment configuration.
•Setting up a scheduled batch job (Code Engine's cron equivalent) for periodic data processing that doesn't justify a full Kubernetes cluster.
•Wiring a Code Engine application to an event source (Cloud Object Storage events, Kafka topic) so it runs only when work arrives.
•Migrating a small Kubernetes deployment to Code Engine when the operational overhead of Kubernetes is hard to justify for the workload size.

Pro Tips

TIP

Set min-scale to 1 for latency-sensitive HTTP apps to avoid cold starts; leave it at 0 for tolerant workloads to get the full pay-per-use benefit. The trade-off is concrete: 1 instance always-on costs roughly $5-15/month, vs. potentially $0 with cold-start latency under load.

TIP

Use secrets and config maps instead of environment variables for anything sensitive. Code Engine treats secrets specially in logs and the dashboard, while env vars can show up in CLI output and incident reports.

How It Works Under the Hood

The builder collects project name, region, and a list of components (apps, jobs, secrets, event subscriptions). For each app/job it captures the image, run command, min/max scale, resource limits, environment variables, and triggers. Output is `ibm_code_engine_project` and component-specific Terraform resources (`ibm_code_engine_app`, `ibm_code_engine_job`, `ibm_code_engine_secret`).

Frequently Asked Questions

When should I choose Code Engine over IKS?: Code Engine is right when your workload is essentially stateless containers without significant per-pod customization (no DaemonSets, no node-level configuration, no PersistentVolume needs). IKS is right when you need Kubernetes-native primitives, complex multi-service routing, custom operators, or persistent storage at scale. For 'I have a web app and a worker', Code Engine is dramatically simpler.
How does Code Engine pricing work?: You pay per second of execution for applications and jobs, plus a small per-1000-request fee. There's a generous free tier each month. Idle applications scale to zero and stop billing. This means low-traffic apps cost almost nothing, while high-throughput apps cost more than a comparably-sized always-on Droplet would — model both before committing.

Was this tool helpful?

Disclaimer: This tool runs entirely in your browser. No data is sent to our servers. Always verify outputs before using them in production. AWS, Azure, and GCP are trademarks of their respective owners.