AWS Fargate module

This module creates a set of AWS resources:

• AWS ECS Fargate Service
• AWS Security group for the Fargate Service
• A set of monitoring resources including Slack notifications
• A set of autoscaling resources
• If autoscaling isn't enabled, desires count will match min and max count

Autoscaling

The autoscaling is triggered by a set of cloudwatch alarms that monitor CPU and memory. Monitoring period and thresholds for each alarm are configurable. Autoscaling resources have been configured using Cloudposse ecs-cloudwatch-autoscaling module

Both scaling up and down are configured to make adjustments by adding or removing a single tasks.

Scaling-up cool down period is 1 minute while scaling-down cool down period is 5 minutes.

Inputs

Name	Description	Type	Default	Required
alarm_data_missing_action	Missing data action for success responses alarm. Possible values: missing or breaching	string	"missing"	no
assign_public_ip	Assign public IP to the Fargate service	bool	false	no
cpu_utilization_high_period	Duration of the monitoring period	number	300	no
cpu_utilization_high_threshold	High CPU threshold	number	60	no
cpu_utilization_low_period	Duration of the monitoring period	number	300	no
cpu_utilization_low_threshold	Low CPU threshold	number	20	no
create_connection_error_alarm	Set to true if connection error alarm should be created	bool	false	no
create_request_count_alarm	Set to true if request count alarm should be created	bool	false	no
create_success_responses_alarm	Set to true if success responses alarm should be created	bool	false	no
create_target_response_time_alarm	Set to true if target response time alarm should be created	bool	false	no
create_unhealthy_host_count_alarm	Set to true if unhealthy host count alarm should be created	bool	false	no
deployment_controller	Type of deployment controller. Valid values: CODE_DEPLOY, ECS, EXTERNAL. Default: ECS	string	"ECS"	no
desired_count	desired number of container instances running	number	n/a	yes
ecs_cluster_arn	ECS Cluster ARN	string	n/a	yes
ecs_cluster_name	Name of the ECS cluster	string	n/a	yes
enable_autoscaling	Indicate if autoscaling should be enabled or not	bool	false	no
enable_codedeploy_control	Setting this variable to true configures Fargate service terraform lifecycle to ignore changes done to the task definition and load balancer config. These will be controlled by code deploy.	bool	false	no
enable_slack_notifications	Indicates if slack notifications should be enabled or not. If true, slack_webhook_url must be provided.	bool	false	no
fargate_platform_version	The version of the Fargate platform	string	"LATEST"	no
health_check_grace_period	Number of seconds that ECS service scheduler should ignore unhealthy ELB target/container/route 53 health checks after a task enters a RUNNING state	number	0	no
include_region	If set to true the current providers region will be appended to any global AWS resources such as IAM roles	bool	false	no
ingress_sg_list	List of ingress security groups that will be applied to the created Fargate service	list(string)	[]	no
max_count	Maximum number of tasks in the service, used only when autoscaling is enabled	number	1	no
max_percent	max percent of healthy container instances	number	200	no
memory_utilization_high_period	Duration of the monitoring period	number	300	no
memory_utilization_high_threshold	High memory threshold	number	60	no
memory_utilization_low_period	Duration of the monitoring period	number	300	no
memory_utilization_low_threshold	Low memory threshold	number	20	no
min_count	Minimum number of tasks in the service, used only when autoscaling is enabled	number	1	no
min_healthy_percent	min percent of healthy container instances	number	100	no
monit_resp_success_percentage	What percentage of requests should be responded to with 2xx	string	"99"	no
monit_target_response_time	service response time in seconds greater than or equal to	string	"0.5"	no
monitoring_config	n/a	list(object({ load_balancer_arn_suffix = string target_group_arn_suffix = string // some of the defaulted properties, such as monitoring period, can be added here }))	n/a	yes
name	Name of the fargate instance	string	"fargate"	no
namespace	n/a	string	n/a	yes
port	The port the service is available from	number	5060	no
region	AWS Region the Fargate service is deployed to	string	n/a	yes
sg_egress_cidr_blocks	List of egress CIDR blocks that will be applied to the created Fargate service	list(string)	[ "0.0.0.0/0" ]	no
sg_list	List of security groups that will be applied to the created Fargate service	list(string)	[]	no
slack_webhook_url	Slack webhook URL for Cloudwatch alarm notifications	string	""	no
stage	Deployment stage i.e. environment name	string	n/a	yes
subnet_ids	A list of subnet ids the fargate service will be deployed to	list(string)	n/a	yes
tags	A set of tags that will be applied to all resources created by this module	map(string)	n/a	yes
target_group_arns	A list of target group ARNs	list(string)	[]	no
task_definition	The family and revision (family:revision) or full ARN of the task definition that you want to run in your service	string	n/a	yes
vpc_id	VPC Id	string	n/a	yes
wait_for_steady_state	Terraform will wait for the service to reach a steady state (like aws ecs wait services-stable) before continuing	bool	false	no
low_cpu_alarm_enabled	"Indicates if the low cpu alarm is enabled	bool	true	yes

Outputs

Name	Description
fargate_service_name	n/a
security_group_id	n/a
sns_alarm_topic_arn	n/a
sns_slack_notification_topic_arn	n/a

Verify Module changes locally.

In order to validate any changes in this repo locally.

1. Push your code up to a brach on BitBucket.
2. Clone the sr-fargate-module repository.
3. in the sr-fargate-module repository open the main.tf file
4. Navigate to line 198 and change the source property for the module "fargate-service" to point to your branch by changing the value of the ref= part of the url to be like ref=BranchName.
5. Pull the latest terraform code from the branch by running the command of terraform init.