UPD: added monitoring for MSK/Redis

modules/ManagementGovernance/Monitoring.Kafka/main.tf

@@ -20,3 +20,127 @@ resource "aws_cloudwatch_metric_alarm" "Kafka-ZooKeeperRequestLatencyMsMean" {
     ignore_changes = [tags]
   }
 }
+
+data "aws_msk_cluster" "msk-cluster" {
+  cluster_name = var.cluster-name
+}
+
+data "aws_msk_broker_nodes" "msk-broker" {
+  cluster_arn = data.aws_msk_cluster.msk-cluster.arn
+}
+
+/*
+output debug {
+  value = data.aws_msk_broker_nodes.msk-broker.node_info_list
+}
+*/
+/*
+module "msk-brokers" {
+  source        = "../../util/resource-list"
+  resource-type = "kafka-brokers"
+  query-input   = data.aws_msk_cluster.msk-cluster.arn
+}
+*/
+
+resource "aws_cloudwatch_metric_alarm" "Kafka-CpuUserSystem" {
+  for_each                  = toset([for i in data.aws_msk_broker_nodes.msk-broker.node_info_list[*].broker_id : tostring(i)])
+  alarm_name                = "${var.cw-alarm-prefix}:Kafka:CpuUsage:${var.cluster-name}-${each.value}"
+  comparison_operator       = "GreaterThanThreshold"
+  evaluation_periods        = "3"
+  threshold                 = var.threshold-CpuUserSystem
+  alarm_description         = "Kafka:ZooKeeperRequestLatencyMsMean"
+  insufficient_data_actions = []
+  actions_enabled           = var.actions-enabled
+  alarm_actions             = [var.sns-targets.alarm-actions-urgent]
+  ok_actions                = [var.sns-targets.alarm-actions-urgent]
+  metric_query {
+    id = "m1"
+    metric {
+      metric_name = "CpuUser"
+      namespace   = "AWS/Kafka"
+      period      = 300
+      stat        = "Average"
+      dimensions = {
+        "Cluster Name" = var.cluster-name
+        "Broker ID"    = each.value
+      }
+    }
+  }
+
+  metric_query {
+    id = "m2"
+    metric {
+      metric_name = "CpuSystem"
+      namespace   = "AWS/Kafka"
+      period      = 300
+      stat        = "Average"
+      dimensions = {
+        "Cluster Name" = var.cluster-name
+        "Broker ID"    = each.value
+      }
+    }
+  }
+
+  metric_query {
+    id          = "e1"
+    expression  = "m1 + m2"
+    label       = "CpuUserSystem"
+    return_data = "true"
+  }
+
+  tags = var.default-tags
+  lifecycle {
+    ignore_changes = [tags]
+  }
+}
+
+resource "aws_cloudwatch_metric_alarm" "Kafka-KafkaDataLogsDiskUsed" {
+  for_each                  = toset([for i in data.aws_msk_broker_nodes.msk-broker.node_info_list[*].broker_id : tostring(i)])
+  alarm_name                = "${var.cw-alarm-prefix}:Kafka:KafkaDataLogsDiskUsed:${var.cluster-name}-${each.value}"
+  comparison_operator       = "GreaterThanThreshold"
+  evaluation_periods        = "3"
+  metric_name               = "KafkaDataLogsDiskUsed"
+  period                    = "300"
+  statistic                 = "Average"
+  threshold                 = var.threshold-KafkaDataLogsDiskUsed
+  alarm_description         = "Kafka:KafkaDataLogsDiskUsed"
+  namespace                 = "AWS/Kafka"
+  insufficient_data_actions = []
+  actions_enabled           = var.actions-enabled
+  alarm_actions             = [var.sns-targets.alarm-actions-urgent]
+  ok_actions                = [var.sns-targets.alarm-actions-urgent]
+  dimensions = {
+    "Cluster Name" = var.cluster-name
+    "Broker ID"    = each.value
+  }
+  tags = var.default-tags
+  lifecycle {
+    ignore_changes = [tags]
+  }
+}
+
+resource "aws_cloudwatch_metric_alarm" "Kafka-HeapMemoryAfterGC" {
+  for_each                  = toset([for i in data.aws_msk_broker_nodes.msk-broker.node_info_list[*].broker_id : tostring(i)])
+  alarm_name                = "${var.cw-alarm-prefix}:Kafka:HeapMemoryAfterGC:${var.cluster-name}-${each.value}"
+  comparison_operator       = "GreaterThanThreshold"
+  evaluation_periods        = "3"
+  metric_name               = "HeapMemoryAfterGC"
+  period                    = "300"
+  statistic                 = "Average"
+  threshold                 = var.threshold-HeapMemoryAfterGC
+  alarm_description         = "Kafka:HeapMemoryAfterGC"
+  namespace                 = "AWS/Kafka"
+  insufficient_data_actions = []
+  actions_enabled           = var.actions-enabled
+  alarm_actions             = [var.sns-targets.alarm-actions-urgent]
+  ok_actions                = [var.sns-targets.alarm-actions-urgent]
+  dimensions = {
+    "Cluster Name" = var.cluster-name
+    "Broker ID"    = each.value
+  }
+  tags = var.default-tags
+  lifecycle {
+    ignore_changes = [tags]
+  }
+}
+

modules/ManagementGovernance/Monitoring.Kafka/variables.tf

@@ -5,3 +5,6 @@ variable sns-targets {}
 variable default-tags {}
 
 variable threshold-ZooKeeperRequestLatencyMsMean {}
+variable threshold-CpuUserSystem {}
+variable threshold-KafkaDataLogsDiskUsed {}
+variable threshold-HeapMemoryAfterGC {}

modules/ManagementGovernance/Monitoring.Redis/main.tf

@@ -66,4 +66,28 @@ resource "aws_cloudwatch_metric_alarm" "redis-CacheHitRate" {
   lifecycle {
     ignore_changes = [tags]
   }
+}
+
+resource "aws_cloudwatch_metric_alarm" "redis-StringBasedCmdsLatency" {
+  alarm_name                = "${var.cw-alarm-prefix}:Redis:StringBasedCmdsLatency:${var.redis-cluster-id}"
+  comparison_operator       = "GreaterThanThreshold"
+  evaluation_periods        = "3"
+  metric_name               = "StringBasedCmdsLatency"
+  period                    = "60"
+  statistic                 = "Average"
+  threshold                 = var.threshold-StringBasedCmdsLatency
+  alarm_description         = "The average latency, in microseconds, of the string-based commands run during a selected time range"
+  namespace                 = "AWS/ElastiCache"
+  insufficient_data_actions = []
+  actions_enabled           = var.actions-enabled
+  alarm_actions             = [var.sns-targets.alarm-actions-standard]
+  ok_actions                = [var.sns-targets.alarm-actions-standard]
+  dimensions = {
+    CacheClusterId = var.redis-cluster-id
+    # CacheNodeId = each.value
+  }
+  tags = var.default-tags
+  lifecycle {
+    ignore_changes = [tags]
+  }
 }

modules/ManagementGovernance/Monitoring.Redis/variables.tf

@@ -1,9 +1,10 @@
-variable cw-alarm-prefix {}
+variable "cw-alarm-prefix" {}
 variable "actions-enabled" {}
 variable "redis-cluster-id" {}
-variable sns-targets {}
+variable "sns-targets" {}
 variable "default-tags" {}
 
 variable "threshold-EngineCPUUtilization" {}
 variable "threshold-DatabaseMemoryUsagePercentage" {}
-variable "threshold-CacheHitRate" {}
+variable "threshold-CacheHitRate" {}
+variable "threshold-StringBasedCmdsLatency" {}

modules/networking/vpc_subnets/README.md

@@ -10,19 +10,18 @@ This module performs the following tasks:
 
 Subnet cidrs are calculated automatically. Due to the design of terraform's cidrsubnets, this module has limitations: 
 
-* supports 2, 4, 6, or 8 subnets in total. 
+* supports 2, 4, 6, 8, or 12 subnets in total. 
 * hard-coded to work with 2 AZs, regardless of number of AZs available in the region. 
 
 Based on the input variables, it will create subnet cidrs using the following function
 
-| Private Subnets per az | Public Subnets per az | Function                                     | Example if a /24 is used on VPC |
+| Private Subnets per az | Public Subnets per az | Function                                             | Example if a /24 is used on VPC |
-| ---------------------- | --------------------- | -------------------------------------------- | ------------------------------- |
+|------------------------|-----------------------|------------------------------------------------------|---------------------------------|
-| 1                      | 0                     | cidrsubnets(local.vpc-cidr, 1,1)             | 2 * /25                         |
+| 1                      | 0                     | cidrsubnets(local.vpc-cidr, 1,1)                     | 2 * /25                         |
-| 1                      | 1                     | cidrsubnets(local.vpc-cidr, 2,2,2,2)         | 4 * /26                         |
+| 1                      | 1                     | cidrsubnets(local.vpc-cidr, 2,2,2,2)                 | 4 * /26                         |
-| 2                      | 1                     | cidrsubnets(local.vpc-cidr, 3,3,3,3,3,3)     | 6 * /27                         |
+| 2                      | 1                     | cidrsubnets(local.vpc-cidr, 3,3,3,3,3,3)             | 6 * /27                         |
-| 2                      | 2                     | cidrsubnets(local.vpc-cidr, 4,4,4,4,4,4,4,4) | 8 * /28                         |
+| 2                      | 2                     | cidrsubnets(local.vpc-cidr, 3,3,3,3,3,3,3,3)         | 8 * /27                         |
-
+| 6                      | 6                     | cidrsubnets(local.vpc-cidr, 4,4,4,4,4,4,4,4,4,4,4,4) | 12 * /28                        |
-simple-divide = local.total-no-subnets >=8 ? cidrsubnets(local.vpc-cidr, 4,4,4,4,4,4,4,4) : local.total-no-subnets >=6 ? cidrsubnets(local.vpc-cidr, 3,3,3,3,3,3) : local.total-no-subnets >=4 ? cidrsubnets(local.vpc-cidr, 2,2,2,2) : local.total-no-subnets >=2 ? cidrsubnets(local.vpc-cidr, 1,1) : null
 
 ## Inputs:
 

modules/networking/vpc_subnets/vpc.tf

@@ -9,7 +9,8 @@ locals {
   vpc-cidr = var.vpc-cidr
   total-no-subnets = local.no-az * (var.number-of-private-subnets-per-az + var.number-of-public-subnets-per-az)
 
-  simple-divide = local.total-no-subnets >=8 ? cidrsubnets(local.vpc-cidr, 4,4,4,4,4,4,4,4) : local.total-no-subnets >=6 ? cidrsubnets(local.vpc-cidr, 3,3,3,3,3,3) : local.total-no-subnets >=4 ? cidrsubnets(local.vpc-cidr, 2,2,2,2) : local.total-no-subnets >=2 ? cidrsubnets(local.vpc-cidr, 1,1) : null
+  # simple-divide = local.total-no-subnets >=8 ? cidrsubnets(local.vpc-cidr, 4,4,4,4,4,4,4,4) : local.total-no-subnets >=6 ? cidrsubnets(local.vpc-cidr, 3,3,3,3,3,3) : local.total-no-subnets >=4 ? cidrsubnets(local.vpc-cidr, 2,2,2,2) : local.total-no-subnets >=2 ? cidrsubnets(local.vpc-cidr, 1,1) : null
+  simple-divide   = local.total-no-subnets >= 12 ? cidrsubnets(local.vpc-cidr, 4,4,4,4, 4,4,4,4, 4,4,4,4) : local.total-no-subnets >= 8 ? cidrsubnets(local.vpc-cidr, 3,3,3,3, 3,3,3,3) : local.total-no-subnets >= 6 ? cidrsubnets(local.vpc-cidr, 3,3,3, 3,3,3) : local.total-no-subnets >= 4 ? cidrsubnets(local.vpc-cidr, 2,2,2,2) : local.total-no-subnets >= 2 ? cidrsubnets(local.vpc-cidr, 1, 1) : null
   public-subnets = slice(local.simple-divide, 0, var.number-of-public-subnets-per-az * local.no-az)
   private-subnets = slice(local.simple-divide, var.number-of-public-subnets-per-az * local.no-az , local.total-no-subnets)
 }