Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.orbitra.atomo.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
Getting started with Orbitra Flows

Orbitra Flows is your workflow orchestration framework built on top of Prefect and integrated with Orbitra’s environment. It provides a streamlined way to build, deploy, and monitor data pipelines and automated workflows in your organization’s infrastructure.

​
What is Prefect?

Prefect is a modern workflow orchestration platform that allows you to build, observe, and react to data pipelines. Think of it as a way to define Python functions as tasks, chain them together into workflows (called “flows”), and execute them reliably with features like retries, scheduling, and observability.

​
Flows and Tasks

In Prefect terminology:
  • Task: A Python function decorated with @task that represents a discrete unit of work
  • Flow: A Python function decorated with @flow that orchestrates multiple tasks into a complete workflow
Tasks can be retried, cached, and monitored independently. Flows provide the overall structure and can be scheduled, triggered, or executed on-demand. Important: When using Orbitra Flows, the @orbitra_deployment decorator replaces the Prefect @flow decorator. You don’t need both - @orbitra_deployment handles flow registration and deployment configuration in one step.

​
Orbitra Flows vs Plain Prefect

While you can use Prefect directly, Orbitra Flows provides:
  • Declarative Deployments: Use the @orbitra_deployment decorator to register and manage flow deployments consistently:
    • Multiple schedules with cron expressions, intervals, and RRULE support
    • Custom container sizing and concurrency limits for your deployments
    • Pre-configured connections to Orbitra Lake, authentication, and compute resources
    • Managed execution identity integrated with the Orbitra ecosystem (Lake, Flows, Email, and related services)
    • Monitoring and alerting integrations, including notifications via Microsoft Teams and Slack
    • Self-hosted infrastructure alerting for operational visibility
    • Multiple pre-configured self-hosted worker profiles (Container Instances Spot, dedicated Container Instances, and VMs)
    • Repository templates with pre-configured CI/CD pipelines for flow deployment
    • Cloud storage for retry logic to reuse results if task/flow fails

​
Orbitra Flows SDK in 2 minutes

​
1) Define a simple flow

from prefect import task
from orbitra.flows.orbitra_deploy import orbitra_deployment

@task
def extract_data():
    # Simulate data extraction
    return {"records": [1, 2, 3, 4, 5]}

@task
def transform_data(data):
    # Simulate transformation
    return {"records": [x * 2 for x in data["records"]]}

@task
def load_data(data):
    # Simulate loading
    print(f"Loading {len(data['records'])} records")
    return data

@orbitra_deployment(name="simple-etl")
def simple_etl_flow():
    raw_data = extract_data()
    transformed_data = transform_data(raw_data)
    result = load_data(transformed_data)
    return result

# Run locally
if __name__ == "__main__":
    simple_etl_flow()

​
2) Add retry logic and result persistence

Orbitra Flows works seamlessly with Orbitra Lake for data persistence. The persist_result=True parameter enables cloud storage of task results, allowing subsequent runs to reuse successful task outputs if the flow fails: 
from prefect import task
from orbitra.flows.orbitra_deploy import orbitra_deployment
from orbitra.lake import get_lake_client
import pandas as pd

@task(retries=3, retry_delay_seconds=15, persist_result=True)
def fetch_and_store_data():
    # Your data fetching logic
    df = pd.DataFrame({
        "id": [1, 2, 3],
        "value": [100, 200, 300],
        "date": ["2026-01-10", "2026-01-11", "2026-01-12"]
    })
    
    # Store in Orbitra Lake
    client = get_lake_client()
    client.overwrite_data(
        namespace="analytics",
        table_name="daily_metrics",
        df=df
    )
    return len(df)

@orbitra_deployment(name="data-ingestion")
def data_ingestion_flow():
    records_processed = fetch_and_store_data()
    print(f"Successfully processed {records_processed} records")

if __name__ == "__main__":
    data_ingestion_flow()

​
3) Add production configuration

Now let’s add schedules, resource limits, and other production settings: 
from prefect import task
from orbitra.flows.orbitra_deploy import orbitra_deployment
from orbitra.lake import get_lake_client
import pandas as pd

@task
def fetch_and_store_data():
    df = pd.DataFrame({
        "id": [1, 2, 3],
        "value": [100, 200, 300],
        "date": ["2026-01-10", "2026-01-11", "2026-01-12"]
    })
    client = get_lake_client()
    client.overwrite_data(
        namespace="analytics",
        table_name="daily_metrics",
        df=df
    )
    return len(df)

@orbitra_deployment(
    name="daily-data-ingestion",
    description="Ingests daily metrics into the analytics namespace",
    schedules=["cron=0 2 * * *"],  # Run daily at 2 AM
    tags=["analytics", "daily"],
    concurrency=1,  # Only allow 1 concurrent run
    container_size="S",  # 2 GB memory / 1 CPU
    enable_schedules_on_creation=True
)
def data_ingestion_flow():
    records_processed = fetch_and_store_data()
    print(f"Successfully processed {records_processed} records")

if __name__ == "__main__":
    data_ingestion_flow()
The @orbitra_deployment decorator handles flow definition and deployment configuration. Here we’ve added:
  • A cron schedule to run daily at 2 AM
  • Concurrency limit to prevent overlapping runs
  • Tags for organization
  • Container size configuration
  • Enabled schedules on creation

​
Advanced Configuration

The @orbitra_deployment decorator supports advanced options for container sizes if you must run heavy computation. You can use predefined sizes (“XS”, “S”, “M”, “L”) or specify custom resources: 
from orbitra.flows.orbitra_deploy import orbitra_deployment, ContainerSize

@orbitra_deployment(
    name="advanced-etl",
    description="Production ETL with custom resources",
    container_size=ContainerSize(memory_gb=4, cpu_cores=2),  # Custom resources
)
def advanced_flow():
    pass

​
Monitoring and Observability

Orbitra Flows automatically tracks:
  • Flow run status (success, failure, running);
  • Task execution times and states;
  • Logs from all tasks and flows;
  • Retry attempts and failures.
Access the Prefect UI to monitor your workflows in real-time. All flows deployed via Orbitra Flows are automatically connected to your organization’s Prefect server workspace.

​
Need more details?

For complete deployment and configuration options, see: