feat: Added auto-scaling functionality to the consumer and redoer.

.github/actions/setup-opentofu/action.yaml

@@ -32,12 +32,13 @@ runs:
       run: |
         variables=(
           "apply_database_updates_immediately" "consumer_container_count"
-          "consumer_cpu" "consumer_memory" "database_instance_count"
+          "consumer_container_max" "consumer_cpu" "consumer_memory"
+          "consumer_message_threshold" "database_instance_count"
           "database_skip_final_snapshot" "deletion_protection"
           "deployment_environments" "environment" "export_expiration"
           "image_tags_mutable" "key_recovery_period" "log_level" "program"
-          "project" "redoer_container_count" "redoer_cpu" "redoer_memory"
-          "region" "repository"
+          "project" "queue_empty_threshold" "redoer_container_count"
+          "redoer_cpu" "redoer_memory" "region" "repository"
         )
         for var in ${variables[@]}; do
           name="TF_VAR_$(echo $var | tr '[:lower:]' '[:upper:]')"

.github/workflows/deploy.yaml

@@ -42,8 +42,10 @@ jobs:
       AWS_ROLE_ARN: ${{ secrets.AWS_ROLE_ARN }}
       TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY: ${{ secrets.TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY }}
       TF_VAR_CONSUMER_CONTAINER_COUNT: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_COUNT }}
+      TF_VAR_CONSUMER_CONTAINER_MAX: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_MAX }}
       TF_VAR_CONSUMER_CPU: ${{ secrets.TF_VAR_CONSUMER_CPU }}
       TF_VAR_CONSUMER_MEMORY: ${{ secrets.TF_VAR_CONSUMER_MEMORY }}
+      TF_VAR_CONSUMER_MESSAGE_THRESHOLD: ${{ secrets.TF_VAR_CONSUMER_MESSAGE_THRESHOLD }}
       TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT: ${{ secrets.TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT }}
       TF_VAR_DELETION_PROTECTION: ${{ secrets.TF_VAR_DELETION_PROTECTION }}
       TF_VAR_DEPLOYMENT_ENVIRONMENTS: ${{ secrets.TF_VAR_DEPLOYMENT_ENVIRONMENTS }}
@@ -54,6 +56,7 @@ jobs:
       TF_VAR_LOG_LEVEL: ${{ secrets.TF_VAR_LOG_LEVEL }}
       TF_VAR_PROGRAM: ${{ secrets.TF_VAR_PROGRAM }}
       TF_VAR_PROJECT: ${{ secrets.TF_VAR_PROJECT }}
+      TF_VAR_QUEUE_EMPTY_THRESHOLD: ${{ secrets.TF_VAR_QUEUE_EMPTY_THRESHOLD }}
       TF_VAR_REDOER_CONTAINER_COUNT: ${{ secrets.TF_VAR_REDOER_CONTAINER_COUNT }}
       TF_VAR_REDOER_CPU: ${{ secrets.TF_VAR_REDOER_CPU }}
       TF_VAR_REDOER_MEMORY: ${{ secrets.TF_VAR_REDOER_MEMORY }}
@@ -95,8 +98,10 @@ jobs:
         env:
           TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY: ${{ secrets.TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY }}
           TF_VAR_CONSUMER_CONTAINER_COUNT: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_COUNT }}
+          TF_VAR_CONSUMER_CONTAINER_MAX: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_MAX }}
           TF_VAR_CONSUMER_CPU: ${{ secrets.TF_VAR_CONSUMER_CPU }}
           TF_VAR_CONSUMER_MEMORY: ${{ secrets.TF_VAR_CONSUMER_MEMORY }}
+          TF_VAR_CONSUMER_MESSAGE_THRESHOLD: ${{ secrets.TF_VAR_CONSUMER_MESSAGE_THRESHOLD }}
           TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT: ${{ secrets.TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT }}
           TF_VAR_DATABASE_INSTANCE_COUNT: ${{ secrets.TF_VAR_DATABASE_INSTANCE_COUNT }}
           TF_VAR_DELETION_PROTECTION: ${{ secrets.TF_VAR_DELETION_PROTECTION }}
@@ -108,6 +113,7 @@ jobs:
           TF_VAR_LOG_LEVEL: ${{ secrets.TF_VAR_LOG_LEVEL }}
           TF_VAR_PROJECT: ${{ secrets.TF_VAR_PROJECT }}
           TF_VAR_PROGRAM: ${{ secrets.TF_VAR_PROGRAM }}
+          TF_VAR_QUEUE_EMPTY_THRESHOLD: ${{ secrets.TF_VAR_QUEUE_EMPTY_THRESHOLD }}
           TF_VAR_REDOER_CONTAINER_COUNT: ${{ secrets.TF_VAR_REDOER_CONTAINER_COUNT }}
           TF_VAR_REDOER_CPU: ${{ secrets.TF_VAR_REDOER_CPU }}
           TF_VAR_REDOER_MEMORY: ${{ secrets.TF_VAR_REDOER_MEMORY }}

.github/workflows/export.yaml

@@ -38,8 +38,10 @@ jobs:
         env:
           TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY: ${{ secrets.TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY }}
           TF_VAR_CONSUMER_CONTAINER_COUNT: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_COUNT }}
+          TF_VAR_CONSUMER_CONTAINER_MAX: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_MAX }}
           TF_VAR_CONSUMER_CPU: ${{ secrets.TF_VAR_CONSUMER_CPU }}
           TF_VAR_CONSUMER_MEMORY: ${{ secrets.TF_VAR_CONSUMER_MEMORY }}
+          TF_VAR_CONSUMER_MESSAGE_THRESHOLD: ${{ secrets.TF_VAR_CONSUMER_MESSAGE_THRESHOLD }}
           TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT: ${{ secrets.TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT }}
           TF_VAR_DATABASE_INSTANCE_COUNT: ${{ secrets.TF_VAR_DATABASE_INSTANCE_COUNT }}
           TF_VAR_DELETION_PROTECTION: ${{ secrets.TF_VAR_DELETION_PROTECTION }}
@@ -51,6 +53,7 @@ jobs:
           TF_VAR_LOG_LEVEL: ${{ secrets.TF_VAR_LOG_LEVEL }}
           TF_VAR_PROJECT: ${{ secrets.TF_VAR_PROJECT }}
           TF_VAR_PROGRAM: ${{ secrets.TF_VAR_PROGRAM }}
+          TF_VAR_QUEUE_EMPTY_THRESHOLD: ${{ secrets.TF_VAR_QUEUE_EMPTY_THRESHOLD }}
           TF_VAR_REPO_OIDC_ARN: ${{ secrets.TF_VAR_REPO_OIDC_ARN }}
           TF_VAR_REPOSITORY: ${{ secrets.TF_VAR_REPOSITORY }}
           TF_VAR_VPC_CIDR: ${{ secrets.TF_VAR_VPC_CIDR }}

.github/workflows/launch-tools.yaml

@@ -46,8 +46,10 @@ jobs:
         env:
           TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY: ${{ secrets.TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY }}
           TF_VAR_CONSUMER_CONTAINER_COUNT: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_COUNT }}
+          TF_VAR_CONSUMER_CONTAINER_MAX: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_MAX }}
           TF_VAR_CONSUMER_CPU: ${{ secrets.TF_VAR_CONSUMER_CPU }}
           TF_VAR_CONSUMER_MEMORY: ${{ secrets.TF_VAR_CONSUMER_MEMORY }}
+          TF_VAR_CONSUMER_MESSAGE_THRESHOLD: ${{ secrets.TF_VAR_CONSUMER_MESSAGE_THRESHOLD }}
           TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT: ${{ secrets.TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT }}
           TF_VAR_DATABASE_INSTANCE_COUNT: ${{ secrets.TF_VAR_DATABASE_INSTANCE_COUNT }}
           TF_VAR_DELETION_PROTECTION: ${{ secrets.TF_VAR_DELETION_PROTECTION }}
@@ -59,6 +61,7 @@ jobs:
           TF_VAR_LOG_LEVEL: ${{ secrets.TF_VAR_LOG_LEVEL }}
           TF_VAR_PROJECT: ${{ secrets.TF_VAR_PROJECT }}
           TF_VAR_PROGRAM: ${{ secrets.TF_VAR_PROGRAM }}
+          TF_VAR_QUEUE_EMPTY_THRESHOLD: ${{ secrets.TF_VAR_QUEUE_EMPTY_THRESHOLD }}
           TF_VAR_REPO_OIDC_ARN: ${{ secrets.TF_VAR_REPO_OIDC_ARN }}
           TF_VAR_REPOSITORY: ${{ secrets.TF_VAR_REPOSITORY }}
           TF_VAR_VPC_CIDR: ${{ secrets.TF_VAR_VPC_CIDR }}

.github/workflows/plan.yaml

@@ -30,10 +30,14 @@ on:
         required: false
       TF_VAR_CONSUMER_CONTAINER_COUNT:
         required: false
+      TF_VAR_CONSUMER_CONTAINER_MAX:
+        required: false
       TF_VAR_CONSUMER_CPU:
         required: false
       TF_VAR_CONSUMER_MEMORY:
         required: false
+      TF_VAR_CONSUMER_MESSAGE_THRESHOLD:
+        required: false
       TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT:
         required: false
       TF_VAR_DELETION_PROTECTION:
@@ -54,6 +58,8 @@ on:
         required: false
       TF_VAR_PROJECT:
         required: false
+      TF_VAR_QUEUE_EMPTY_THRESHOLD:
+        required: false
       TF_VAR_REDOER_CONTAINER_COUNT:
         required: false
       TF_VAR_REDOER_CPU:
@@ -113,8 +119,10 @@ jobs:
         env:
           TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY: ${{ secrets.TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY }}
           TF_VAR_CONSUMER_CONTAINER_COUNT: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_COUNT }}
+          TF_VAR_CONSUMER_CONTAINER_MAX: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_MAX }}
           TF_VAR_CONSUMER_CPU: ${{ secrets.TF_VAR_CONSUMER_CPU }}
           TF_VAR_CONSUMER_MEMORY: ${{ secrets.TF_VAR_CONSUMER_MEMORY }}
+          TF_VAR_CONSUMER_MESSAGE_THRESHOLD: ${{ secrets.TF_VAR_CONSUMER_MESSAGE_THRESHOLD }}
           TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT: ${{ secrets.TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT }}
           TF_VAR_DATABASE_INSTANCE_COUNT: ${{ secrets.TF_VAR_DATABASE_INSTANCE_COUNT }}
           TF_VAR_DELETION_PROTECTION: ${{ secrets.TF_VAR_DELETION_PROTECTION }}
@@ -126,6 +134,7 @@ jobs:
           TF_VAR_LOG_LEVEL: ${{ secrets.TF_VAR_LOG_LEVEL }}
           TF_VAR_PROJECT: ${{ secrets.TF_VAR_PROJECT }}
           TF_VAR_PROGRAM: ${{ secrets.TF_VAR_PROGRAM }}
+          TF_VAR_QUEUE_EMPTY_THRESHOLD: ${{ secrets.TF_VAR_QUEUE_EMPTY_THRESHOLD }}
           TF_VAR_REDOER_CONTAINER_COUNT: ${{ secrets.TF_VAR_REDOER_CONTAINER_COUNT }}
           TF_VAR_REDOER_CPU: ${{ secrets.TF_VAR_REDOER_CPU }}
           TF_VAR_REDOER_MEMORY: ${{ secrets.TF_VAR_REDOER_MEMORY }}

.github/workflows/pull-request.yaml

@@ -49,15 +49,18 @@ jobs:
       TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY: ${{ secrets.TF_VAR_APPLY_DATABASE_UPDATES_IMMEDIATELY }}
       TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT: ${{ secrets.TF_VAR_DATABASE_SKIP_FINAL_SNAPSHOT }}
       TF_VAR_CONSUMER_CONTAINER_COUNT: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_COUNT }}
+      TF_VAR_CONSUMER_CONTAINER_MAX: ${{ secrets.TF_VAR_CONSUMER_CONTAINER_MAX }}
       TF_VAR_CONSUMER_CPU: ${{ secrets.TF_VAR_CONSUMER_CPU }}
       TF_VAR_CONSUMER_MEMORY: ${{ secrets.TF_VAR_CONSUMER_MEMORY }}
+      TF_VAR_CONSUMER_MESSAGE_THRESHOLD: ${{ secrets.TF_VAR_CONSUMER_MESSAGE_THRESHOLD }}
       TF_VAR_DELETION_PROTECTION: ${{ secrets.TF_VAR_DELETION_PROTECTION }}
       TF_VAR_DEPLOYMENT_ENVIRONMENTS: ${{ secrets.TF_VAR_DEPLOYMENT_ENVIRONMENTS }}
       TF_VAR_EXPORT_EXPIRATION: ${{ secrets.TF_VAR_EXPORT_EXPIRATION }}
       TF_VAR_IMAGE_TAGS_MUTABLE: ${{ secrets.TF_VAR_IMAGE_TAGS_MUTABLE }}
       TF_VAR_KEY_RECOVERY_PERIOD: ${{ secrets.TF_VAR_KEY_RECOVERY_PERIOD }}
       TF_VAR_LOG_LEVEL: ${{ secrets.TF_VAR_LOG_LEVEL }}
       TF_VAR_PROGRAM: ${{ secrets.TF_VAR_PROGRAM }}
+      TF_VAR_QUEUE_EMPTY_THRESHOLD: ${{ secrets.TF_VAR_QUEUE_EMPTY_THRESHOLD }}
       TF_VAR_REDOER_CONTAINER_COUNT: ${{ secrets.TF_VAR_REDOER_CONTAINER_COUNT }}
       TF_VAR_REDOER_CPU: ${{ secrets.TF_VAR_REDOER_CPU }}
       TF_VAR_REDOER_MEMORY: ${{ secrets.TF_VAR_REDOER_MEMORY }}

README.md

@@ -119,6 +119,13 @@ sz_command -C add_record \
   PEOPLE 1 '{"NAME_FULL":"Robert Smith", "DATE_OF_BIRTH":"7/4/1976", "PHONE_NUMBER":"555-555-2088"}'
 ```
 
+You can also run commands using the tools container without entering an
+interactive shell. For example, to purge the repository, you could run:
+
+```bash
+docker compose run tools sz_command -C "purge_repository --FORCEPURGE"
+```
+
 #### Loading sample data
 
 From inside the tools container:

tofu/config/service/main.tf

@@ -18,15 +18,16 @@ module "inputs" {
 module "system" {
   source = "../../modules/system"
 
-  environment         = var.environment
-  project             = var.project
-  export_expiration   = var.export_expiration
-  key_recovery_period = var.key_recovery_period
-  logging_bucket      = module.inputs.values["logging/bucket"]
-  logging_key_arn     = module.inputs.values["logging/key"]
-  log_level           = var.log_level
-  tags                = merge({ awsApplication : module.inputs.values["application/tag"] }, var.tags)
-  vpc_id              = module.inputs.values["vpc/id"]
+  environment           = var.environment
+  project               = var.project
+  export_expiration     = var.export_expiration
+  key_recovery_period   = var.key_recovery_period
+  logging_bucket        = module.inputs.values["logging/bucket"]
+  logging_key_arn       = module.inputs.values["logging/key"]
+  log_level             = var.log_level
+  tags                  = merge({ awsApplication : module.inputs.values["application/tag"] }, var.tags)
+  vpc_id                = module.inputs.values["vpc/id"]
+  queue_empty_threshold = var.queue_empty_threshold
 
   database_subnets                   = split(",", module.inputs.values["vpc/private_subnets"])
   apply_database_updates_immediately = var.apply_database_updates_immediately
@@ -36,11 +37,13 @@ module "system" {
   image_tag                          = local.image_tag
   image_tags_mutable                 = var.image_tags_mutable
 
-  container_subnets        = split(",", module.inputs.values["vpc/private_subnets"])
-  consumer_container_count = var.consumer_container_count
-  consumer_cpu             = var.consumer_cpu
-  consumer_memory          = var.consumer_memory
-  redoer_container_count   = var.redoer_container_count
-  redoer_cpu               = var.redoer_cpu
-  redoer_memory            = var.redoer_memory
+  container_subnets          = split(",", module.inputs.values["vpc/private_subnets"])
+  consumer_container_count   = var.consumer_container_count
+  consumer_container_max     = var.consumer_container_max
+  consumer_cpu               = var.consumer_cpu
+  consumer_memory            = var.consumer_memory
+  consumer_message_threshold = var.consumer_message_threshold
+  redoer_container_count     = var.redoer_container_count
+  redoer_cpu                 = var.redoer_cpu
+  redoer_memory              = var.redoer_memory
 }

tofu/config/service/variables.tf

@@ -10,6 +10,12 @@ variable "consumer_container_count" {
   default     = 1
 }
 
+variable "consumer_container_max" {
+  type        = number
+  description = "Maximum number of consumer containers to run."
+  default     = 10
+}
+
 variable "consumer_cpu" {
   type        = number
   description = "Number of virtual CPUs to allocate to each consumer container."
@@ -22,6 +28,12 @@ variable "consumer_memory" {
   default     = 4096
 }
 
+variable "consumer_message_threshold" {
+  type        = number
+  description = "Number of messages in the SQS queue that will trigger scaling up the number of consumer containers."
+  default     = 250000
+}
+
 variable "database_instance_count" {
   type        = number
   description = "Number of instances in the database cluster."
@@ -103,6 +115,12 @@ variable "project" {
   default     = "sqs-senzing"
 }
 
+variable "queue_empty_threshold" {
+  type        = number
+  description = "Number of minutes that the SQS queue must have zero messages before we consider it empty."
+  default     = 15
+}
+
 variable "redoer_container_count" {
   type        = number
   description = "Desired number of redoer containers to run."

tofu/modules/persistent_service/locals.tf

@@ -0,0 +1,3 @@
+locals {
+  prefix = join("-", compact([var.project, var.environment, var.service]))
+}

tofu/modules/persistent_service/main.tf

@@ -29,11 +29,11 @@ module "service" {
   source  = "HENNGE/ecs/aws//modules/core/service"
   version = "5.3.0"
 
-  cluster                = var.cluster_arn
-  name                   = join("-", compact([var.project, var.environment, var.service]))
-  create_task_definition = false
-  task_definition_arn    = module.task.task_definition_arn
-  desired_count          = var.desired_containers
+  cluster                      = var.cluster_name
+  name                         = local.prefix
+  create_task_definition       = false
+  task_definition_arn          = module.task.task_definition_arn
+  ignore_desired_count_changes = true
 
   launch_type                   = "FARGATE"
   task_requires_compatibilities = ["FARGATE"]
@@ -48,3 +48,13 @@ module "service" {
 
   tags = var.tags
 }
+
+module "scaling_target" {
+  source  = "HENNGE/ecs/aws//modules/core/ecs-autoscaling-target"
+  version = "5.3.0"
+
+  ecs_cluster_name = var.cluster_name
+  ecs_service_name = module.service.name
+  min_capacity     = var.desired_containers
+  max_capacity     = var.max_containers
+}

tofu/modules/persistent_service/outputs.tf

@@ -13,6 +13,16 @@ output "service_name" {
   value       = module.service.name
 }
 
+output "scale_down_policy_arn" {
+  description = "ARN of the scale-down auto-scaling policy."
+  value       = var.scale_down_policy.enabled ? aws_appautoscaling_policy.down["this"].arn : null
+}
+
+output "scale_up_policy_arn" {
+  description = "ARN of the scale-up auto-scaling policy."
+  value       = var.scale_up_policy.enabled ? aws_appautoscaling_policy.up["this"].arn : null
+}
+
 output "task_definition_arn" {
   description = "ARN of the ECS task definition."
   value       = module.task.task_definition_arn

tofu/modules/persistent_service/scaling.tf

@@ -0,0 +1,52 @@
+resource "aws_appautoscaling_policy" "up" {
+  for_each = var.scale_up_policy.enabled ? toset(["this"]) : toset([])
+
+  name               = "${local.prefix}-up"
+  policy_type        = "StepScaling"
+  resource_id        = module.scaling_target.resource_id
+  scalable_dimension = module.scaling_target.scalable_dimension
+  service_namespace  = module.scaling_target.service_namespace
+
+  step_scaling_policy_configuration {
+    adjustment_type = "ExactCapacity"
+    cooldown        = 60
+
+    dynamic "step_adjustment" {
+      for_each = range(1, var.max_containers + 1)
+
+      content {
+        metric_interval_lower_bound = (step_adjustment.value == 1
+          ? var.scale_up_policy.start - 1
+          : (step_adjustment.value - 1) * var.scale_up_policy.step + 1
+        )
+        metric_interval_upper_bound = (step_adjustment.value == var.max_containers
+          ? null
+          : step_adjustment.value * var.scale_up_policy.step + 1
+        )
+        scaling_adjustment = step_adjustment.value
+      }
+    }
+  }
+}
+
+resource "aws_appautoscaling_policy" "down" {
+  for_each = var.scale_down_policy.enabled ? toset(["this"]) : toset([])
+
+  name               = "${local.prefix}-down"
+  policy_type        = "StepScaling"
+  resource_id        = module.scaling_target.resource_id
+  scalable_dimension = module.scaling_target.scalable_dimension
+  service_namespace  = module.scaling_target.service_namespace
+
+  step_scaling_policy_configuration {
+    # We've already waited for the queue to be empty for a threshold of time, so
+    # we don't need a big delay here.
+    adjustment_type = "ExactCapacity"
+    cooldown        = 60
+
+    step_adjustment {
+      metric_interval_upper_bound = 0
+      scaling_adjustment          = 0
+    }
+  }
+}