AWS Managed Services for Production Workloads

Getting onto AWS is the easy part. Keeping an account tidy, secure and affordable over years, while the application changes underneath it, is where most teams struggle. With AWS managed services, an engineer who knows your environment is responsible for it.

We work with the services production applications actually use: EC2, ECS and Fargate, ECR, RDS, S3, CloudFront, load balancers, VPC networking, IAM and CloudWatch. Case Study 1 describes a Fargate environment we built with Terraform and CodePipeline.

What managed means, month to month

Infrastructure as code, not console clicks

Environments built by hand drift. Environments defined in Terraform can be reviewed, repeated and rebuilt. Here is a trimmed example of the kind of definition we use for a private Fargate service.

resource "aws_ecs_service" "backend" {
  name            = "backend"
  cluster         = aws_ecs_cluster.main.id
  task_definition = aws_ecs_task_definition.backend.arn
  desired_count   = 2

  capacity_provider_strategy {
    capacity_provider = "FARGATE"
    weight            = 1
  }

  network_configuration {
    subnets          = var.private_subnet_ids
    security_groups  = [aws_security_group.backend.id]
    assign_public_ip = false
  }

  # load_balancer { ... } attaches the service to the ALB target group
}