Skip to main content

​
orbitra.flows.orbitra_deploy

​
Functions

​
get_registered_deployments

get_registered_deployments() -> list[OrbitraDeploymentMeta]

​
orbitra_deployment

orbitra_deployment(name: Optional[str] = None, description: Optional[str] = None, concurrency: Optional[int] = 1, schedules: Optional[List[str]] = None, enable_schedules_on_creation: bool = False, tags: Optional[List[str]] = None, flow_config: Callable[..., PrefectFlow] = prefect_flow(log_prints=True), schedules_timezone: Optional[str] = None, container_size: ContainerSizeType = None, collision_strategy: Optional[Literal['ENQUEUE', 'CANCEL_NEW']] = 'CANCEL_NEW') -> Callable[[Callable[..., Any]], PrefectFlow]
Decorator to register an Orbitra deployment. This decorator records metadata of a flow so it can later be deployed. It also automatically transforms the decorated function into a Prefect Flow (equivalent to @flow). Args:
  • name: A friendly name for the deployment. Defaults to the function’s name if not provided.
  • description: A free-form description of the deployment. Can be used to document the purpose, ownership, or context of the flow.
  • concurrency: Maximum number of concurrent runs allowed for this deployment. If omitted, no concurrency limit is enforced.
  • schedules: A list of schedule definitions attached to the deployment. Supported formats include
    • "cron={expr}" for cron-based schedules.
    • "interval={seconds}" for fixed interval schedules.
    • Any valid RRULE string (RFC 5545) for advanced recurrence rules.
  • enable_schedules_on_creation: If True, schedules are created on active state for new deployments. Useful for defining schedules and enabling them by default.
  • tags: A list of tags to associate with the deployment. Tags can be used for organization, filtering, or triggering rules.
  • flow_config: A callable returned by prefect.flow(...). The decorated function will be wrapped with it.
  • schedules_timezone: Timezone name to apply to all schedules for this deployment (e.g., "America/Sao_Paulo"). If omitted, "America/Sao_Paulo" will be used as the default timezone.
  • container_size: Defines the container resources to be used when generating the deployment infrastructure. Supported presets are "XS"(default), "S", "M", and "L", which map to
    • XS -> 1 GB memory / 1 CPU
    • S -> 2 GB memory / 1 CPU
    • M -> 4 GB memory / 2 CPUs
    • L -> 8 GB memory / 4 CPUs
Pass ContainerSize(memory_gb=..., cpu_cores=...) to define bespoke resources. Both memory_gb and cpu_cores must be integers; floats are not accepted. Values must also stay within the Azure Container Apps limits (1-240 GB RAM, 1-31 vCPU). If not defined, the default for the work pool will be used.
  • collision_strategy: Strategy to use when a new run is started while another is already running.
Examples: 
from prefect import flow
from orbitra.flows.orbitra_deploy import orbitra_deployment, ContainerSize

@orbitra_deployment()
def manual_flow():
    ...

@orbitra_deployment(
    name="reporting-prod",
    flow_config=flow(name="daily-reporting", log_prints=True),
    schedules=["cron=0 7 * * *"],
    tags=["prod", "finance"],
    container_size="M",
)
def generate_reports(day: str = "yesterday"):
    ...

@orbitra_deployment(
    flow_config=flow(name="etl-runner", retries=2, retry_delay_seconds=60),
    schedules=["interval=3600"],
    container_size=ContainerSize(memory_gb=16, cpu_cores=6),
)
def hourly_etl():
    ...

​
Classes

​
ContainerSize

​
OrbitraDeploymentMeta