AWS Auto Scaling Groups Builder
AWS Auto Scaling group is a great way of managing Amazon EC2 instances. AWS Auto Scaling group watches the corresponding instances' health and launches new instances whenever needed: to replace instances that became unhealthy or to scale the group. Furthermore, with an AWS Elastic Load Balancer attached, AWS Auto Scaling Group also takes ELB health checks into account, making it possible to replace the instances based on service-level signals.
However, creating AWS Auto Scaling Groups might be tricky. To create a new group, one needs to perform many actions using either AWS Management Console or AWS CLI. In addition, the user needs to wait until the corresponding resources become available for some steps. It might not be easy to recall all the steps required if preparing new services is not a regular job.
So, this is a job for an automation tool, such as AWS Auto Scaling Groups Builder. It uses AWS EC2 API, AWS EC2 Auto Scaling API, and AWS ELB API to perform the following operations automatically:
- register an AMI from a running instance;
- create a launch template using this AMI;
- register a target group;
- create a load balancer with a listener that forwards traffic to this target group;
- create an Auto Scaling group with both EC2 and ELB health checks;
- handle all the errors that might happen along the way; in case of failure, all the created artifacts are deleted.
The tool uses the default AWS credentials config. Run aws configure or set up the environment variables AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, and AWS_REGION before running the tool.
Only AWS EC2 Auto Scaling is supported. The service is supposed to handle health checks via HTTP.
The tool offers some default settings for all the services, which are not available for changing via command-line arguments. However, it's always possible to change them manually with the default AWS tooling after the service is ready. These defaults are:
- All the instances of the created service inherit the following settings of the source instance:
- instance type;
- kernel ID;
- key name;
- licence specifications;
- placement;
- network interfaces.
- The service only uses spot instances, despite the source instance lifecycle option.
- The tool creates and AMI with name
$GROUP_NAME v1. The launch template name is just$GROUP_NAME. The names of target group and balancer are both$GROUP_NAMEwith substitution:-instead of_. E.g., if you're going to create an Auto Scaling group with namemy_service, the following resources will be created:- AMI with
my_service v1; - launch template
my_service; - target group
my-service; - load balancer
my-service; - Auto Scaling group
my_service.
- AMI with
- The balancer is created in all availability zones, in the default subnets. So, if one has six availability zones in their AWS account, all these zones will be used for placing the load balancer.
- The instances for the service are also placed within the default subnets of all the availability zones of the AWS account.
- The Auto Scaling group has the following capacity settings:
maximum capacity = 2 * desired capacity = 2 * minimum capacity;- capacity rebalancing is turned on;
- no dynamic or predictive scaling policies, no scheduled actions.
Examples of Use
The following command will create an Auto Scaling group my_service_group from an instance i-0699803d818227e16:
aws_asg_builder --instance i-0699803d818227e16 --group my_service_group --port 8080 --instances 10
The output looks like that then:
2022/01/07 10:42:01 will create an AMI "my_service_group v1" from the instance i-0699803d818227e16
2022/01/07 10:42:01 will create a launch template "my_service_group"
2022/01/07 10:42:01 will create a target group "my-service-group" in VPC vpc-c832bdb5
2022/01/07 10:42:01 will create a load balancer "my-service-group" in subnets subnet-7e353733, subnet-bb0d39b5, subnet-6bafd034, subnet-6b7c0e0d, subnet-66ca4b57, subnet-dd8ef7fc
2022/01/07 10:42:01 will create an auto scaling group "my_service_group" with 10 t2.small spot instances
2022/01/07 10:42:01 ami-00c78efe33a162b36 ("my_service_group v1"): pending
2022/01/07 10:43:02 ami-00c78efe33a162b36 ("my_service_group v1"): pending
2022/01/07 10:44:02 ami-00c78efe33a162b36 ("my_service_group v1"): pending
2022/01/07 10:45:02 ami-00c78efe33a162b36 ("my_service_group v1"): available
2022/01/07 10:45:02 created launch template "my_service_group"
2022/01/07 10:45:03 created target group "my-service-group" (arn:aws:elasticloadbalancing:us-east-1:046898261394:targetgroup/my-service-group/9ff35ebad2a86a0d)
2022/01/07 10:45:05 load balancer "my-service-group": provisioning
2022/01/07 10:46:05 load balancer "my-service-group": provisioning
2022/01/07 10:47:05 load balancer "my-service-group": active
2022/01/07 10:47:09 group "my_service_group": 0 instances in total, 0 instances are in service and healthy (10 needed)
2022/01/07 10:48:09 group "my_service_group", instance "i-0071e23d5f723ca78": InService, Healthy
2022/01/07 10:48:09 group "my_service_group", instance "i-0250bc8b1ef1173d2": InService, Healthy
2022/01/07 10:48:09 group "my_service_group", instance "i-02c51ec311822e072": InService, Healthy
2022/01/07 10:48:09 group "my_service_group", instance "i-0415988f40e7cb87e": InService, Healthy
2022/01/07 10:48:09 group "my_service_group", instance "i-0470f10de375fc978": Pending, Healthy
2022/01/07 10:48:09 group "my_service_group", instance "i-047cd43b77947afce": InService, Healthy
2022/01/07 10:48:09 group "my_service_group", instance "i-087d57ef33e6c26b0": Pending, Healthy
2022/01/07 10:48:09 group "my_service_group", instance "i-08d426cade3c3bf8e": Pending, Healthy
2022/01/07 10:48:09 group "my_service_group", instance "i-09a68825c258faf9f": InService, Healthy
2022/01/07 10:48:09 group "my_service_group", instance "i-0bb9b9061e7970004": InService, Healthy
2022/01/07 10:48:09 group "my_service_group": 10 instances in total, 7 instances are in service and healthy (10 needed)
2022/01/07 10:49:10 group "my_service_group", instance "i-0071e23d5f723ca78": InService, Healthy
2022/01/07 10:49:10 group "my_service_group", instance "i-0250bc8b1ef1173d2": InService, Healthy
2022/01/07 10:49:10 group "my_service_group", instance "i-02c51ec311822e072": InService, Healthy
2022/01/07 10:49:10 group "my_service_group", instance "i-0415988f40e7cb87e": InService, Healthy
2022/01/07 10:49:10 group "my_service_group", instance "i-0470f10de375fc978": InService, Healthy
2022/01/07 10:49:10 group "my_service_group", instance "i-047cd43b77947afce": InService, Healthy
2022/01/07 10:49:10 group "my_service_group", instance "i-087d57ef33e6c26b0": InService, Healthy
2022/01/07 10:49:10 group "my_service_group", instance "i-08d426cade3c3bf8e": InService, Healthy
2022/01/07 10:49:10 group "my_service_group", instance "i-09a68825c258faf9f": InService, Healthy
2022/01/07 10:49:10 group "my_service_group", instance "i-0bb9b9061e7970004": InService, Healthy
2022/01/07 10:49:10 group "my_service_group": 10 instances in total, 10 instances are in service and healthy (10 needed)
2022/01/07 10:49:10 successfully created an auto scaling group "my_service_group"
2022/01/07 10:49:10 enabled metrics collection for the group "my_service_group"
2022/01/07 10:49:10 AMI link: https://console.aws.amazon.com/ec2/v2/home?region=us-east-1#ImageDetails:imageId=ami-00c78efe33a162b36
2022/01/07 10:49:10 Launch template link: https://console.aws.amazon.com/ec2/v2/home?region=us-east-1#LaunchTemplateDetails:launchTemplateId=lt-0ce496d3c72c69e9d
2022/01/07 10:49:10 Target group link: https://console.aws.amazon.com/ec2/v2/home?region=us-east-1#TargetGroup:targetGroupArn=arn:aws:elasticloadbalancing:us-east-1:046898261394:targetgroup/my-service-group/9ff35ebad2a86a0d
2022/01/07 10:49:10 Balancer link: https://console.aws.amazon.com/ec2/v2/home?region=us-east-1#LoadBalancers:search=my-service-group
2022/01/07 10:49:10 Auto Scalingr group link: https://console.aws.amazon.com/ec2autoscaling/home?region=us-east-1#/details/my_service_group
2022/01/07 10:49:10 check out the health status: http://my-service-group-1643996769.us-east-1.elb.amazonaws.com:8080/health
Program Arguments
group: the name of the Auto Scaling group to create; required.instance: AWS EC2 instance ID to create the service from; required.health-path: the health HTTP handler for the service; optional, default:/health.port: the HTTP traffic port for the service; optional, default:80.instances: the number of instances to create within the group; optional, default:1.health-check-grace-period: the time needed for the instance to become healthy after the launch; optional, default:1m. Use the Golang duration strings to override, see https://pkg.go.dev/time#ParseDuration.update-timeout: the time limit to complete the instance refresh; optional, default:30m. Use the Golang duration strings to override, see https://pkg.go.dev/time#ParseDuration.update-tick: the time between status updates in the log file; optional, default:1m. Making this parameter lower might speed up the overall execution. Use the Golang duration strings to override, see https://pkg.go.dev/time#ParseDuration.
Installation
Assuming you already have Golang installed on the machine, simply run:
go install github.com/ashagraev/aws_asg_builder@latest
The tool will then appear in your golang binary folder (e.g., ~/go/bin/aws_asg_builder). If you don't have Golang yet, consider installing it using the official guide https://go.dev/doc/install.
Alternatively, you can download the pre-built binaries from the latest release: https://github.com/ashagraev/aws_asg_builder/releases/latest.
