​

cogni.yaml Reference

​

Basic Structure

version: 1
start:
  command: ["python", "server.py"]
  port: 8000
env:
  - name: ENV_VAR
    value: "value"
resources:
  cpu: 1
  memory: 512Mi

​

Options

​

version

• Type: integer
• Required: Yes
• Description: Version of the configuration schema. Use 1 for current deployments.

​

start

• Type: object
• Required: Yes
• Description: How to start your server.
  • command: List of strings. The command to run your server (e.g., ["python", "server.py"]).
  • port: Integer. The port your server listens on (e.g., 8000).

​

env

• Type: list of objects
• Required: No
• Description: Environment variables to set in the container.
  • name: Name of the environment variable.
  • value: Value to set.

​

resources

• Type: object
• Required: No
• Description: Resource limits for your server.
  • cpu: Number of CPUs to allocate (e.g., 1)
  • memory: Amount of memory to allocate (e.g., 512Mi)

​

Advanced Options

• healthcheck: Command or HTTP endpoint to check server health
• volumes: List of volume mounts
• args: Additional arguments to pass to the start command

​

Example: Full Configuration

version: 1
start:
  command: ["python", "main.py"]
  port: 8080
env:
  - name: ENV
    value: "production"
  - name: API_KEY
    value: "your-key"
resources:
  cpu: 2
  memory: 1Gi
healthcheck:
  http: "/healthz"
volumes:
  - name: data
    mountPath: /app/data

​

Best Practices

• Validation: Use a schema or linter to validate your cogni.yaml before deploying.
• Secrets: Do not store sensitive secrets directly in cogni.yaml. Use environment variables or secret management tools.
• Port: Ensure the port matches the port your server actually listens on.
• Minimalism: Only specify options you need; defaults are used for unspecified fields.

​

Troubleshooting

• If deployment fails, check for YAML syntax errors and ensure all required fields are present.
• Use local testing to verify your configuration before deploying to production.