Back to docs

Stack Configuration Reference

Reference for .ai-dlc/settings.yml stack layers — infrastructure, compute, packaging, pipeline, secrets, monitoring, alerting, and operations

The stack configuration describes your project's infrastructure in .ai-dlc/settings.yml under the stack: key. AI-DLC uses this configuration to generate deployment manifests, validate operations, and tailor advice to your environment.

All layers are optional. An empty stack (stack: {}) is valid — AI-DLC will infer what it can from project files.

Layer Reference

Infrastructure

Defines how infrastructure is provisioned.

stack:
  infrastructure:
    - provider: terraform  # terraform | cloudformation | pulumi
      state_backend: s3
      modules_path: infra/modules/

Providers: terraform, cloudformation, pulumi

Compute

Defines where workloads run.

stack:
  compute:
    - provider: kubernetes  # kubernetes | ecs | lambda | docker-compose
      cluster: production
      namespace: default

Providers: kubernetes, ecs, lambda, docker-compose

Packaging

Defines how applications are packaged for deployment.

stack:
  packaging:
    - provider: helm  # helm | kustomize | raw
      charts_path: deploy/charts/

Providers: helm, kustomize, raw

Pipeline

Defines CI/CD pipeline configuration.

stack:
  pipeline:
    - provider: github-actions  # github-actions | gitlab-ci | jenkins | circleci
      workflows_path: .github/workflows/

Providers: github-actions, gitlab-ci, jenkins, circleci

Secrets

Defines secrets management.

stack:
  secrets:
    - provider: vault  # vault | aws-sm | gcp-sm | env
      path: secret/data/app

Providers: vault, aws-sm, gcp-sm, env

Monitoring

Defines observability stack.

stack:
  monitoring:
    - provider: prometheus  # prometheus | datadog | cloudwatch | newrelic | otel
      endpoint: http://prometheus:9090

Providers: prometheus, datadog, cloudwatch, newrelic, otel

Alerting

Defines alerting and incident management.

stack:
  alerting:
    - provider: pagerduty  # pagerduty | opsgenie | datadog | alertmanager
      service_id: P123ABC

Providers: pagerduty, opsgenie, datadog, alertmanager

Operations

Defines the operations runtime and configuration.

stack:
  operations:
    runtime: node  # node | python | go | shell
    scheduled:
      - name: cleanup
        schedule: "0 3 * * *"
        command: node scripts/cleanup.js
    reactive:
      - name: alert-handler
        trigger: webhook
        command: node scripts/handle-alert.js

Runtimes: node, python, go, shell

When runtime is not specified, AI-DLC auto-detects from project files:

  • package.json → node
  • pyproject.toml or requirements.txt → python
  • go.mod → go
  • Default → shell

Examples

Simple — Solo Developer

A minimal setup with just CI and basic monitoring:

stack:
  pipeline:
    - provider: github-actions
  monitoring:
    - provider: datadog

Medium — Small Team

A typical team setup with container orchestration, CI, monitoring, and operations:

stack:
  compute:
    - provider: kubernetes
      cluster: staging
      namespace: app
  pipeline:
    - provider: github-actions
      workflows_path: .github/workflows/
  monitoring:
    - provider: prometheus
      endpoint: http://prometheus:9090
  alerting:
    - provider: pagerduty
      service_id: P456DEF
  operations:
    runtime: node
    scheduled:
      - name: cleanup
        schedule: "0 3 * * *"
        command: node scripts/cleanup.js

Complex — Enterprise

A fully populated stack for a large-scale production environment:

stack:
  infrastructure:
    - provider: terraform
      state_backend: s3
      modules_path: infra/modules/
  compute:
    - provider: kubernetes
      cluster: prod-us-east-1
      namespace: platform
  packaging:
    - provider: helm
      charts_path: deploy/charts/
  pipeline:
    - provider: github-actions
      workflows_path: .github/workflows/
  secrets:
    - provider: vault
      path: secret/data/platform
  monitoring:
    - provider: otel
      endpoint: http://otel-collector:4317
  alerting:
    - provider: pagerduty
      service_id: P789GHI
  operations:
    runtime: node
    scheduled:
      - name: cleanup
        schedule: "0 3 * * *"
        command: node scripts/cleanup.js
    reactive:
      - name: alert-handler
        trigger: webhook
        command: node scripts/handle-alert.js

Next Steps