Configuration overview

reconD reads one YAML config file at startup. The packaged service uses /etc/recond/config.yaml; pass -config to use another path (see CLI).

• The file must have a .yaml or .yml extension.
• It must define a top-level workloads list, even for a single workload.
• Workload names must be unique, because the name keys the workload's deployment state.
• Changes take effect only after a restart: sudo systemctl restart recond.

Structure

log:                 # optional: process logging
  format: json
  level: info

deployment_retry:    # optional: retry defaults for every workload
  max_retry_forward_deployment: 1
  max_retry_rollback: 1

workloads:           # required: one entry per independently deployed service
  - name: my-app
    container_name: my-app
    deploy_command: /opt/my-app/deploy.sh
    git_provider:
      owner: my-github-owner
      repo: my-app
    labels:
      app: my-app   # must match a LABEL on your image; scopes image pruning

Complete example

log:
  format: text
  level: info

deployment_retry:
  max_retry_forward_deployment: 3
  max_retry_rollback: 2
  initial_interval: 1s
  max_interval: 30s
  multiplier: 2

workloads:
  - name: my-app
    release_prefix: my-app-prod
    container_name: my-app
    deploy_command: cd /opt/my-app && make redeploy-app
    check_interval: 60s
    notification_url: https://hooks.slack.com/services/T000/B000/XXXX
    git_provider:
      owner: my-github-owner
      repo: my-deployment-metadata-repo
      token: ""            # falls back to the GITHUB_TOKEN environment variable
    strategy:
      type: env_file
      env_file_path: /opt/my-app/.env.deploy
      image_tag_key: IMAGE_TAG
    health_check:
      type: http
      url: http://localhost:8080/health
      retries: 12
      interval: 10s
      timeout: 5s
    labels:
      app: my-app

  - name: worker
    release_prefix: worker-prod
    container_name: worker
    deploy_command: /opt/worker/deploy.sh
    git_provider:
      owner: my-github-owner
      repo: my-deployment-metadata-repo
    health_check:
      type: tcp
      url: tcp://localhost:9000
    labels:
      app: worker
    deployment_retry:
      max_retry_forward_deployment: 1   # overrides only this field

Durations

Duration fields accept either a Go duration string (500ms, 10s, 1m30s) or a plain integer number of seconds (60).

Validation

reconD validates the whole file before starting any workload and exits with load config error if anything is wrong, naming the workload index and field. For example:

agent error: load config error "/etc/recond/config.yaml": workload 0: deploy_command is required

Where to go next

• Field reference: every field, its default, and its rules.
• Deploy command: how your command is run.
• Deployment strategies, Health checks, Retries, Notifications.