ecs-deployer-boto3

Deploy applications to AWS ECS using boto3 and pydantic.

This library provides a thin, typed wrapper over the AWS ECS deployment API. It is decoupled from how you discover or provision your infrastructure (CloudFormation, CDK, Terraform, or by hand) — you describe your services as pydantic models and the library handles registering task definitions, creating or updating services, cleaning up old revisions, and monitoring deployments.

Supports both EC2 (with placement strategy / constraints) and Fargate (with capacity provider strategy and awsvpc network configuration).

Installation

pip install ecs-deployer-boto3

Usage

import boto3

from ecs_deployer_boto3 import (
    ApplicationUpdater,
    Container,
    DeploymentMonitor,
    Service,
    Task,
)

services = [
    Service(
        name='web',
        cluster='my-cluster',
        task_definition=Task(
            family='web',
            containers=[
                Container(
                    name='web',
                    image='123456789.dkr.ecr.us-east-1.amazonaws.com/web:latest',
                    port_mappings=[{'container_port': 8000}],
                ),
            ],
            cpu='256',
            memory='512',
            requires_compatibilities=['FARGATE'],
        ),
        launch_type='FARGATE',
        network_configuration={
            'subnets': ['subnet-abc'],
            'security_groups': ['sg-abc'],
            'assign_public_ip': 'ENABLED',
        },
    ),
]

session = boto3.Session()

updater = ApplicationUpdater(services, session)
updater.update_application(environment={'FOO': 'bar'})

monitor = DeploymentMonitor(services, session)
monitor.monitor(limit_minutes=15)

License

MIT