Scheduling

Any interface can run on a cron schedule instead of (or in addition to) being invoked over HTTP. Add a schedule: block to the interface. Full fields: Schedule reference.

Example

interfaces:
  nightlyDigest:
    schedule:
      cron: "0 6 * * *"        # 5-field cron: min hour day month weekday
      enabled: true
      timezone: "Europe/London" # IANA zone, defaults to UTC
    actions:
      - name: Query
        database: main
        query: SELECT count(*) AS signups FROM users WHERE created_at > now() - interval '1 day'
      - name: Notify
        http:
          url: https://hooks.slack.com/services/XXX
          body: { text: "a|Query::signups| new signups today" }

Retries

Scheduled runs retry independently of action-level retries:

schedule:
  cron: "*/15 * * * *"
  enabled: true
  max_attempts: 3
  retry_backoff_seconds: 60      # first backoff
  retry_backoff_multiplier: 2.0  # doubles each attempt
  max_backoff_seconds: 3600      # cap

note

timezone and cron are validated when the config loads — an invalid zone or expression is rejected. A minimum interval is enforced (minute-level).