feat: Add autoscaling and capacity provider strategy to ECS examples

Rishang · May 29, 2024 · 85c4092 · 85c4092
1 parent 6e03861
commit 85c4092
Show file tree

Hide file tree

Showing 9 changed files with 147 additions and 95 deletions.
diff --git a/README.md b/README.md
@@ -17,7 +17,7 @@
 
 module "fargate" {
   source  = "Rishang/fargate/aws"
-  version = "1.3.0"
+  version = "1.4.3"
 
   EnvironmentName = "test"
 
@@ -26,13 +26,22 @@ module "fargate" {
   service             = "whoami"
   container_port      = 80
   task_definition_arn = module.fargate_task_definition.arn
-  scale_min_capacity           = 3
+  scale_min_capacity  = 3
+  scale_max_capacity  = 10
   
-  fargate_spot        = true
-  
-  # default to true, only set force_spot to false
-  # if you want to use mixed on-demand and spot instances 
-  force_spot        = false
+  # keep 1 FARGATE for each 5 FARGATE_SPOT
+  capacity_provider_strategy = [
+    {
+      base              = 1
+      capacity_provider = "FARGATE"
+      weight            = 1
+    },
+    {
+      base              = 0
+      capacity_provider = "FARGATE_SPOT"
+      weight            = 5
+    }
+  ]
 
   # networking
   assign_public_ip = true
@@ -49,8 +58,25 @@ module "fargate" {
   point_to_r53 = true
 
   # autoscale (optional)
-  memory_scale_target = 60
-  # cpu_scale_target  = 60
+  create_autoscale_target = true
+  cpu_scale_target  = 60
+  # memory_scale_target = 60
+
+  # scheduled scaling (optional)
+  scaling_schedule = [
+    {
+      # Scale count to zero every night at 19:00
+      schedule     = "cron(0 19 * * ? *)"
+      min_capacity = 0
+      max_capacity = 0
+    },
+    {
+      # Scale count to 3 every morning at 7:00
+      schedule     = "cron(0 7 * * ? *)"
+      min_capacity = 3
+      max_capacity = 3
+    }
+  ]
 
   # service discovery (optional)
   enable_discovery = true
@@ -91,19 +117,27 @@ module "fargate" {
 # null are required inputs, 
 # others are optional default values
 
-EnvironmentName                    = null
-assign_public_ip                   = false
+EnvironmentName  = null
+assign_public_ip = false
+capacity_provider_strategy = [{
+  base              = "1"
+  capacity_provider = "FARGATE"
+  weight            = "1"
+  }, {
+  base              = "0"
+  capacity_provider = "FARGATE_SPOT"
+  weight            = "0"
+}]
 cluster                            = null
 container_name                     = ""
 container_port                     = -1
 cpu_scale_target                   = -1
+create_autoscale_target            = false
 deployment_maximum_percent         = 200
 deployment_minimum_healthy_percent = 100
 enable_discovery                   = false
 enable_ecs_managed_tags            = false
-fargate_spot                       = false
 force_new_deployment               = false
-force_spot                         = false
 health_check_interval              = 20
 health_check_matcher               = "200,202"
 health_check_path                  = "/"
@@ -138,16 +172,16 @@ vpc_id                             = ""
 | <a name="input_subnets"></a> [subnets](#input\_subnets) | List of subnets for ecs service | `list(string)` | n/a | yes |
 | <a name="input_task_definition_arn"></a> [task\_definition\_arn](#input\_task\_definition\_arn) | The ARN of the task definition to use for the ECS service | `string` | n/a | yes |
 | <a name="input_assign_public_ip"></a> [assign\_public\_ip](#input\_assign\_public\_ip) | Auto assign public ip for ecs containers | `bool` | `false` | no |
+| <a name="input_capacity_provider_strategy"></a> [capacity\_provider\_strategy](#input\_capacity\_provider\_strategy) | Capacity provider strategy for ecs service here `base` parameter defines the minimum number of tasks that should be launched using the specified capacity provider before considering the weight. `weight` parameter defines the relative percentage of tasks to be launched using the specified capacity provider after the base tasks have been satisfied. | `list(map(any))` | <pre>[<br>  {<br>    "base": 1,<br>    "capacity_provider": "FARGATE",<br>    "weight": 1<br>  },<br>  {<br>    "base": 0,<br>    "capacity_provider": "FARGATE_SPOT",<br>    "weight": 0<br>  }<br>]</pre> | no |
 | <a name="input_container_name"></a> [container\_name](#input\_container\_name) | Required if service name is different than main application container\_name of task defination | `string` | `""` | no |
 | <a name="input_container_port"></a> [container\_port](#input\_container\_port) | container application port | `number` | `-1` | no |
 | <a name="input_cpu_scale_target"></a> [cpu\_scale\_target](#input\_cpu\_scale\_target) | Treshold cpu target value for autoscaling ecs service | `number` | `-1` | no |
+| <a name="input_create_autoscale_target"></a> [create\_autoscale\_target](#input\_create\_autoscale\_target) | Enable to create autoscale for ecs service | `bool` | `false` | no |
 | <a name="input_deployment_maximum_percent"></a> [deployment\_maximum\_percent](#input\_deployment\_maximum\_percent) | Deployment max healthy percent of container count | `number` | `200` | no |
 | <a name="input_deployment_minimum_healthy_percent"></a> [deployment\_minimum\_healthy\_percent](#input\_deployment\_minimum\_healthy\_percent) | Deployment min healthy percent of container count | `number` | `100` | no |
 | <a name="input_enable_discovery"></a> [enable\_discovery](#input\_enable\_discovery) | Enable service discovery, requires `namespace_id` and `container_name` | `bool` | `false` | no |
 | <a name="input_enable_ecs_managed_tags"></a> [enable\_ecs\_managed\_tags](#input\_enable\_ecs\_managed\_tags) | Specifies whether to enable Amazon ECS managed tags for the service. | `bool` | `false` | no |
-| <a name="input_fargate_spot"></a> [fargate\_spot](#input\_fargate\_spot) | Whether to use fargate spot instances or not. | `bool` | `false` | no |
 | <a name="input_force_new_deployment"></a> [force\_new\_deployment](#input\_force\_new\_deployment) | Enable to force a new task deployment of the service | `bool` | `false` | no |
-| <a name="input_force_spot"></a> [force\_spot](#input\_force\_spot) | Set only fargate spot as Capacity provider. | `bool` | `false` | no |
 | <a name="input_health_check_interval"></a> [health\_check\_interval](#input\_health\_check\_interval) | target group health check interval time in sec | `number` | `20` | no |
 | <a name="input_health_check_matcher"></a> [health\_check\_matcher](#input\_health\_check\_matcher) | Service health check response matcher | `string` | `"200,202"` | no |
 | <a name="input_health_check_path"></a> [health\_check\_path](#input\_health\_check\_path) | Health check path for ecs running containers | `string` | `"/"` | no |

diff --git a/examples/ecs-alb-r53-servicediscovery/main.tf b/examples/ecs-alb-r53-servicediscovery/main.tf
@@ -134,8 +134,22 @@ module "fargate" {
   point_to_r53 = true
 
   # autoscale
-  memory_scale_target = 60
-  # cpu_scale_target    = 60
+  create_autoscale_target = true
+  # cpu_scale_target        = 60
+  scaling_schedule = [
+    {
+      # Scale count to zero every night at 19:00
+      schedule     = "cron(0 19 * * ? *)"
+      min_capacity = 0
+      max_capacity = 0
+    },
+    {
+      # Scale count to 3 every morning at 7:00
+      schedule     = "cron(0 7 * * ? *)"
+      min_capacity = 3
+      max_capacity = 3
+    }
+  ]
 
   # service discovery
   enable_discovery = true

diff --git a/examples/ecs-alb-r53/.terraform.lock.hcl b/examples/ecs-alb-r53/.terraform.lock.hcl
diff --git a/examples/ecs-alb-r53/main.tf b/examples/ecs-alb-r53/main.tf
@@ -109,10 +109,20 @@ module "fargate" {
   container_port      = 80
   task_definition_arn = module.fargate_task_definition.arn
   scale_min_capacity  = 2
-  fargate_spot        = true
 
-  # keep non spot capacity_provider too
-  force_spot = false
+  # keep 1 FARGATE for each 5 FARGATE_SPOT
+  capacity_provider_strategy = [
+    {
+      base              = 1
+      capacity_provider = "FARGATE"
+      weight            = 1
+    },
+    {
+      base              = 0
+      capacity_provider = "FARGATE_SPOT"
+      weight            = 5
+    }
+  ]
 
   # networking
   assign_public_ip = true
@@ -129,10 +139,10 @@ module "fargate" {
   point_to_r53 = true
 
   # autoscale
+  create_autoscale_target = true
   # memory_scale_target = 60
   cpu_scale_target = 60
 
-
   tags = {
     Name         = "whoami"
     Version      = "1.0.0"

diff --git a/examples/ecs/.terraform.lock.hcl b/examples/ecs/.terraform.lock.hcl
diff --git a/examples/ecs/main.tf b/examples/ecs/main.tf
@@ -95,6 +95,7 @@ module "fargate" {
   container_port      = 80
   task_definition_arn = module.fargate_task_definition.arn
   scale_min_capacity  = 1
+  scale_max_capacity  = 5
 
   # networking
   assign_public_ip = true

diff --git a/examples/example.md b/examples/example.md
@@ -10,7 +10,7 @@
 
 module "fargate" {
   source  = "Rishang/fargate/aws"
-  version = "1.3.0"
+  version = "1.4.3"
 
   EnvironmentName = "test"
 
@@ -19,13 +19,22 @@ module "fargate" {
   service             = "whoami"
   container_port      = 80
   task_definition_arn = module.fargate_task_definition.arn
-  scale_min_capacity           = 3
+  scale_min_capacity  = 3
+  scale_max_capacity  = 10
   
-  fargate_spot        = true
-  
-  # default to true, only set force_spot to false
-  # if you want to use mixed on-demand and spot instances 
-  force_spot        = false
+  # keep 1 FARGATE for each 5 FARGATE_SPOT
+  capacity_provider_strategy = [
+    {
+      base              = 1
+      capacity_provider = "FARGATE"
+      weight            = 1
+    },
+    {
+      base              = 0
+      capacity_provider = "FARGATE_SPOT"
+      weight            = 5
+    }
+  ]
 
   # networking
   assign_public_ip = true
@@ -42,8 +51,25 @@ module "fargate" {
   point_to_r53 = true
 
   # autoscale (optional)
-  memory_scale_target = 60
-  # cpu_scale_target  = 60
+  create_autoscale_target = true
+  cpu_scale_target  = 60
+  # memory_scale_target = 60
+
+  # scheduled scaling (optional)
+  scaling_schedule = [
+    {
+      # Scale count to zero every night at 19:00
+      schedule     = "cron(0 19 * * ? *)"
+      min_capacity = 0
+      max_capacity = 0
+    },
+    {
+      # Scale count to 3 every morning at 7:00
+      schedule     = "cron(0 7 * * ? *)"
+      min_capacity = 3
+      max_capacity = 10 
+    }
+  ]
 
   # service discovery (optional)
   enable_discovery = true

diff --git a/main.tf b/main.tf
@@ -10,7 +10,7 @@ locals {
 
   # convert : demo.example.com to example.com
   root_domain = join(".", slice(split(".", var.subdomain), 1, length(split(".", var.subdomain))))
-  asg_target  = (var.cpu_scale_target > 0 || var.memory_scale_target > 0) ? 1 : 0
+  asg_target  = var.create_autoscale_target == true ? 1 : 0
 }
 
 # ------------------------ DATA ------------------------------
@@ -140,23 +140,14 @@ resource "aws_ecs_service" "fargate" {
   enable_ecs_managed_tags = var.enable_ecs_managed_tags
   # for stability a 1 dedicated fargate instance and rest spot
   # 1 dedicated per 5 spot
-  dynamic "capacity_provider_strategy" {
-    for_each = toset(var.fargate_spot == true && var.force_spot == false ? ["create"] : [])
-
-    content {
-      base              = 1
-      capacity_provider = "FARGATE"
-      weight            = 1
-    }
-  }
 
   dynamic "capacity_provider_strategy" {
-    for_each = toset(var.fargate_spot == true ? ["create"] : [])
+    for_each = var.capacity_provider_strategy
 
     content {
-      base              = var.fargate_spot == true && var.force_spot == true ? 1 : 0
-      capacity_provider = "FARGATE_SPOT"
-      weight            = 5
+      base              = capacity_provider_strategy.value["base"]
+      capacity_provider = capacity_provider_strategy.value["capacity_provider"]
+      weight            = capacity_provider_strategy.value["weight"]
     }
   }
 
@@ -233,7 +224,7 @@ resource "aws_appautoscaling_target" "scaling" {
 }
 
 resource "aws_appautoscaling_policy" "cpu_scale_up_policy" {
-  count = var.cpu_scale_target > 0 ? 1 : 0
+  count = (var.create_autoscale_target == true && var.cpu_scale_target > 0) ? 1 : 0
 
   name               = "ECSServiceAverageCPUUtilization:${aws_appautoscaling_target.scaling[count.index].resource_id}"
   policy_type        = "TargetTrackingScaling"
@@ -256,7 +247,7 @@ resource "aws_appautoscaling_policy" "cpu_scale_up_policy" {
 # memory base scale
 
 resource "aws_appautoscaling_policy" "memory_scale_up_policy" {
-  count = var.memory_scale_target > 0 ? 1 : 0
+  count = (var.create_autoscale_target == true && var.memory_scale_target > 0) ? 1 : 0
 
   name               = "ECSServiceAverageMemoryUtilization:${aws_appautoscaling_target.scaling[count.index].resource_id}"
   policy_type        = "TargetTrackingScaling"
@@ -278,8 +269,8 @@ resource "aws_appautoscaling_policy" "memory_scale_up_policy" {
 
 # alb based auto scaling 
 
-resource "aws_appautoscaling_policy" "scale_up_policy" {
-  count = var.lb_scale_target > 0 ? 1 : 0
+resource "aws_appautoscaling_policy" "traffic_scale_up_policy" {
+  count = (var.create_autoscale_target == true && var.lb_scale_target > 0) ? 1 : 0
 
   name               = "ALBRequestCountPerTarget:${aws_appautoscaling_target.scaling[count.index].resource_id}"
   policy_type        = "TargetTrackingScaling"