AWS ECS/Fargate

The AWS provider runs Ctrl Plane instances on Amazon ECS with Fargate, providing serverless container hosting without managing EC2 instances.

Status

Interface defined — the package structure exists at provider/aws and implements the provider.Provider interface. Full implementation is in progress.

Configuration

import "github.com/xraph/ctrlplane/provider/aws"

prov, err := aws.New(aws.Config{
    Region:     "us-east-1",
    Cluster:    "ctrlplane-prod",
    SubnetIDs:  []string{"subnet-abc", "subnet-def"},
    SecurityGroups: []string{"sg-123"},
    ExecutionRoleARN: "arn:aws:iam::123:role/ecsTaskExecution",
    TaskRoleARN: "arn:aws:iam::123:role/ecsTask",
})

Field	Env	Default	Description
`Region`	`CP_AWS_REGION`	—	AWS region
`Cluster`	`CP_AWS_ECS_CLUSTER`	—	ECS cluster name or ARN
`SubnetIDs`	`CP_AWS_SUBNET_IDS`	—	VPC subnet IDs for tasks
`SecurityGroups`	`CP_AWS_SECURITY_GROUPS`	—	Security group IDs
`ExecutionRoleARN`	`CP_AWS_EXEC_ROLE_ARN`	—	ECS task execution IAM role
`TaskRoleARN`	`CP_AWS_TASK_ROLE_ARN`	—	IAM role for the running task

Capabilities

Capability	Supported
`provision`	Yes
`deploy`	Yes
`scale`	Yes
`logs`	Via CloudWatch
`exec`	Via ECS Exec
`rolling`	Yes
`autoscale`	Via Application Auto Scaling
`custom_domains`	Via ALB + Route53
`tls`	Via ACM

Resource mapping

Ctrl Plane concept	AWS resource
Instance	ECS Service + Task Definition
Environment variables	Task Definition environment
Ports	Target Group + ALB listener
Resources (CPU/Memory)	Fargate task size
Custom domains	Route53 record + ALB rule
TLS certificates	ACM certificate
Scaling	Application Auto Scaling policy

How it works

Provision creates an ECS Service with a Fargate task definition.
Deploy registers a new task definition revision and updates the service.
Scale updates the service desired count and task resource limits.
Logs streams from CloudWatch Logs using the task log group.
Exec uses ECS Exec to open a session to the running container.

When to use

Teams already on AWS infrastructure
Serverless container workloads without cluster management
Applications that need tight AWS service integration (RDS, SQS, S3)
Environments where managed auto-scaling is required