API Reference

Attributes:

Name (str): Name of the service.

Description (str): Description.

Image (str): Docker image to deploy.

Entrypoint (str): Command to run inside the container.

Environment (str, default: 'dev'): Deployment environment.

Owner, Group, Permissions, RunAs: User and access control.

AppGroup (str, optional): Logical grouping in the UI.

CPU (int, default: 200): CPU allocation.

Memory (int, default: 250): Memory allocation.

HealthCheckUrl (str, default: '/'): Health endpoint.

SpotInstance (bool, default: False): Use spot pricing.

NumContainers (int, default: 1): Replication count.

ScalingConfig (ScalingConfig): Auto-scaling rules.

Port (int, default: 11111): TCP port the service listens on.

GpuEnabled (bool, default: False): Flag for GPU support.

Methods:

save: Save and deploy the service.

start: Start execution.

stop: Stop containers.

remove: Stop and unregister the service.

status: Property that returns job status.

Class Methods:

Service.list_existing - returns a list of all deployed services, which the user has access to.

Validations:

Entrypoint must exist in the container image.

Port must be in the valid range 1–65535.

Usage example:

from dt.scheduler.api import Service

service = Service(Name='Test Service',
                  Description='This is a test Service',
                  Image='Price data processing',
                  Entrypoint='data_pipeline.data_provider',
                  CPU=200,
                  Memory=200)

service.save()
print(f'{service.status=}')
service.stop()
print(f'{service.status=}')
service.start()
print(f'{service.status=}')
service.remove()
print(f'{service.status=}')
service.save()
print(f'{service.status=}')

This will produce the following output:

service.status='running'
service.status='stopped'
service.status='running'
service.status='non_existing'
service.status='running'

After this, your service will be available in the 'Dev Apps' section under the 'Test Service' button.