Revert "Tweak alert thresholds"

This reverts commit d36bb8f.

For some reason the alarms here constantly re-registered without becoming
stable.  Reverted to previous version and the re-registration has now
stopped.

Will try again to tweak thresholds to values that AWS registers OK.

service-infrastructure/alerts/metric_alarms_resource.tf

@@ -7,7 +7,7 @@ resource "aws_cloudwatch_metric_alarm" "ecs_cpu_usage" {
   evaluation_periods  = 1
   metric_name         = "CPUUtilization"
   namespace           = "AWS/ECS"
-  period              = 30
+  period              = 300
   statistic           = "Average"
   threshold           = 95
 
@@ -36,7 +36,7 @@ resource "aws_cloudwatch_metric_alarm" "ecs_memory_usage" {
   evaluation_periods  = 1
   metric_name         = "MemoryUtilization"
   namespace           = "AWS/ECS"
-  period              = 30
+  period              = 300
   statistic           = "Average"
   threshold           = 95
 
@@ -66,7 +66,7 @@ resource "aws_cloudwatch_metric_alarm" "rds_cpu_usage" {
   evaluation_periods  = 1
   metric_name         = "CPUUtilization"
   namespace           = "AWS/RDS"
-  period              = 30
+  period              = 300
   statistic           = "Average"
   threshold           = 95
 
@@ -94,7 +94,7 @@ resource "aws_cloudwatch_metric_alarm" "rds_cluster_cpu_usage" {
   evaluation_periods  = 1
   metric_name         = "CPUUtilization"
   namespace           = "AWS/RDS"
-  period              = 30
+  period              = 300
   statistic           = "Average"
   threshold           = 95
 
@@ -122,9 +122,9 @@ resource "aws_cloudwatch_metric_alarm" "alb_5xx_errors" {
   evaluation_periods  = 1
   metric_name         = "HTTPCode_ELB_5XX_Count"
   namespace           = "AWS/ApplicationELB"
-  period              = 60
+  period              = 300
   statistic           = "Sum"
-  threshold           = 10
+  threshold           = 0
   treat_missing_data  = "notBreaching"
 
   dimensions = {
@@ -148,12 +148,12 @@ resource "aws_cloudwatch_metric_alarm" "alb_4xx_errors" {
 
   alarm_name          = "${each.value}-4xx-errors"
   comparison_operator = "GreaterThanThreshold"
-  evaluation_periods  = 2
+  evaluation_periods  = 1
   metric_name         = "HTTPCode_ELB_4XX_Count"
   namespace           = "AWS/ApplicationELB"
-  period              = 60
+  period              = 300
   statistic           = "Sum"
-  threshold           = 50
+  threshold           = 100
   treat_missing_data  = "notBreaching"
 
   dimensions = {
@@ -177,7 +177,7 @@ resource "aws_cloudwatch_metric_alarm" "fargate_spot_instance_terminated_by_AWS"
   evaluation_periods  = 1
   metric_name         = aws_cloudwatch_log_metric_filter.fargate_spot_instance_terminated_by_AWS_metric.name
   namespace           = "CISBenchmark"
-  period              = 60
+  period              = 300
   statistic           = "Sum"
   threshold           = 0
   treat_missing_data  = "notBreaching"