add alarms for failed lambdas

terraform/main.tf

@@ -3,15 +3,27 @@ provider "aws" {
 }
 
 module "alpr_counts" {
-  module_name = "alpr_counts"
-  source = "./modules/alpr_counts"
+  module_name          = "alpr_counts"
+  source               = "./modules/alpr_counts"
   deflock_stats_bucket = var.deflock_stats_bucket
-  rate = "rate(60 minutes)"
+  rate                 = "rate(60 minutes)"
+  sns_topic_arn        = aws_sns_topic.lambda_alarms.arn
 }
 
 module "alpr_clusters" {
-  module_name = "alpr_clusters"
-  source = "./modules/alpr_clusters"
-  deflock_cdn_bucket = var.deflock_cdn_bucket
-  rate = "rate(1 hour)"
+  module_name          = "alpr_clusters"
+  source               = "./modules/alpr_clusters"
+  deflock_cdn_bucket   = var.deflock_cdn_bucket
+  rate                 = "rate(1 hour)"
+  sns_topic_arn        = aws_sns_topic.lambda_alarms.arn
+}
+
+resource "aws_sns_topic" "lambda_alarms" {
+  name = "lambda_alarms_topic"
+}
+
+resource "aws_sns_topic_subscription" "sms_subscription" {
+  topic_arn = aws_sns_topic.lambda_alarms.arn
+  protocol  = "sms"
+  endpoint  = var.alarm_phone_number
 }

terraform/modules/alpr_clusters/main.tf

@@ -89,3 +89,23 @@ resource "aws_cloudwatch_log_group" "lambda_log_group" {
   name              = "/aws/lambda/${aws_lambda_function.overpass_lambda.function_name}"
   retention_in_days = 14
 }
+
+resource "aws_sns_topic" "lambda_alarms" {
+  name = "${var.module_name}_lambda_alarms"
+}
+
+resource "aws_cloudwatch_metric_alarm" "lambda_error_alarm" {
+  alarm_name          = "${var.module_name}_execution_error"
+  alarm_description   = "An error has occurred while executing the ${var.module_name} Lambda"
+  namespace           = "AWS/Lambda"
+  metric_name         = "Errors"
+  dimensions          = {
+    FunctionName = aws_lambda_function.overpass_lambda.function_name
+  }
+  statistic           = "Sum"
+  period              = 86400 # 1 day
+  evaluation_periods  = 1
+  threshold           = 0
+  comparison_operator = "GreaterThanThreshold"
+  alarm_actions       = [var.sns_topic_arn]
+}

terraform/modules/alpr_clusters/variables.tf

@@ -9,3 +9,8 @@ variable "deflock_cdn_bucket" {
 variable "rate" {
   description = "Rate at which to run the Lambda function"
 }
+
+variable "sns_topic_arn" {
+  description = "The ARN of the SNS topic for Lambda alarms"
+  type        = string
+}

terraform/modules/alpr_counts/main.tf

@@ -120,3 +120,22 @@ resource "aws_iam_role_policy_attachment" "lambda_cloudwatch_logs_attachment" {
   role       = aws_iam_role.lambda_role.name
   policy_arn = aws_iam_policy.lambda_cloudwatch_logs_policy.arn
 }
+
+# Alarms for Failure
+
+resource "aws_cloudwatch_metric_alarm" "lambda_error_alarm" {
+  alarm_name          = "${var.module_name}_execution_error"
+  alarm_description   = "An error has occurred while executing the ${var.module_name} Lambda"
+  comparison_operator = "GreaterThanThreshold"
+  evaluation_periods  = 1
+  metric_name         = "Errors"
+  namespace           = "AWS/Lambda"
+  period              = 86400 # 1 day
+  statistic           = "Sum"
+  threshold           = 0
+  dimensions = {
+    FunctionName = aws_lambda_function.overpass_lambda.function_name
+  }
+
+  alarm_actions = [var.sns_topic_arn]
+}

terraform/modules/alpr_counts/variables.tf

@@ -26,3 +26,8 @@ variable "aws_account_id" {
   type        = string
   default     = "912821578123"
 }
+
+variable "sns_topic_arn" {
+  description = "The ARN of the SNS topic for Lambda alarms"
+  type        = string
+}

terraform/variables.tf

@@ -7,3 +7,8 @@ variable "deflock_cdn_bucket" {
   description = "S3 bucket for the CDN"
   default = "cdn.deflock.me"
 }
+
+variable "alarm_phone_number" {
+  description = "Phone number to receive alarm notifications"
+  # intentionally left blank since this file is checked into git
+}