feat: support config via tags (#6)

marbot-io · May 3, 2023 · 033c44a · 033c44a
1 parent 1069007
commit 033c44a
Show file tree

Hide file tree

Showing 3 changed files with 142 additions and 41 deletions.
diff --git a/README.md b/README.md
@@ -23,6 +23,26 @@ terraform init
 terraform apply
 ```
 
+## Config via tags
+
+You can also configure this module by tagging the RDS cluster (requires v1.0.0 or higher). Tags take precedence over variables (tags override variables).
+
+| tag key                                        | default value                                    | allowed values                                |
+| ---------------------------------------------- | ------------------------------------------------ | ----------------------------------------------|
+| `marbot`                                       | on                                               | on,off                                        |
+| `marbot:cpu-utilization`                       | variable `cpu_utilization`                       | static,off                                    |
+| `marbot:cpu-utilization:threshold`             | variable `cpu_utilization_threshold`             | 0-100                                         |
+| `marbot:cpu-utilization:period`                | variable `cpu_utilization_period`                | <= 86400 and multiple of 60                   |
+| `marbot:cpu-utilization:evaluation-periods`    | variable `cpu_utilization_evaluation_periods`    | >= 1 and $period*$evaluation-periods <= 86400 |
+| `marbot:cpu-credit-balance`                    | variable `cpu_credit_balance`                    | static,off                                    |
+| `marbot:cpu-credit-balance:threshold`          | variable `cpu_credit_balance_threshold`          | >= 0                                          |
+| `marbot:cpu-credit-balance:period`             | variable `cpu_credit_balance_period`             | <= 86400 and multiple of 60                   |
+| `marbot:cpu-credit-balance:evaluation-periods` | variable `cpu_credit_balance_evaluation_periods` | >= 1 and $period*$evaluation-periods <= 86400 |
+| `marbot:freeable-memory`                       | variable `freeable_memory`                       | static,off                                    |
+| `marbot:freeable-memory:threshold`             | variable `freeable_memory_threshold`             | >= 0                                          |
+| `marbot:freeable-memory:period`                | variable `freeable_memory_period`                | <= 86400 and multiple of 60                   |
+| `marbot:freeable-memory:evaluation-periods`    | variable `freeable_memory_evaluation_periods`    | >= 1 and $period*$evaluation-periods <= 86400 |
+
 ## Update procedure
 
 1. Update the `version`

diff --git a/main.tf b/main.tf
@@ -12,8 +12,34 @@ data "aws_caller_identity" "current" {}
 
 data "aws_region" "current" {}
 
+data "aws_rds_cluster" "cluster" {
+  cluster_identifier = var.db_cluster_identifier
+}
+
 locals {
   topic_arn = var.create_topic == false ? var.topic_arn : join("", aws_sns_topic.marbot.*.arn)
+  enabled   = var.enabled && lookup(data.aws_rds_cluster.cluster.tags, "marbot", "on") != "off"
+
+  cpu_utilization                        = lookup(data.aws_rds_cluster.cluster.tags, "marbot:cpu-utilization", var.cpu_utilization)
+  cpu_utilization_threshold              = try(tonumber(lookup(data.aws_rds_cluster.cluster.tags, "marbot:cpu-utilization:threshold", var.cpu_utilization_threshold)), var.cpu_utilization_threshold)
+  cpu_utilization_period_raw             = try(tonumber(lookup(data.aws_rds_cluster.cluster.tags, "marbot:cpu-utilization:period", var.cpu_utilization_period)), var.cpu_utilization_period)
+  cpu_utilization_period                 = min(max(floor(local.cpu_utilization_period_raw / 60) * 60, 60), 86400)
+  cpu_utilization_evaluation_periods_raw = try(tonumber(lookup(data.aws_rds_cluster.cluster.tags, "marbot:cpu-utilization:evaluation-periods", var.cpu_utilization_evaluation_periods)), var.cpu_utilization_evaluation_periods)
+  cpu_utilization_evaluation_periods     = min(max(local.cpu_utilization_evaluation_periods_raw, 1), floor(86400 / local.cpu_utilization_period))
+
+  cpu_credit_balance                        = lookup(data.aws_rds_cluster.cluster.tags, "marbot:cpu-credit-balance", var.cpu_credit_balance)
+  cpu_credit_balance_threshold              = try(tonumber(lookup(data.aws_rds_cluster.cluster.tags, "marbot:cpu-credit-balance:threshold", var.cpu_credit_balance_threshold)), var.cpu_credit_balance_threshold)
+  cpu_credit_balance_period_raw             = try(tonumber(lookup(data.aws_rds_cluster.cluster.tags, "marbot:cpu-credit-balance:period", var.cpu_credit_balance_period)), var.cpu_credit_balance_period)
+  cpu_credit_balance_period                 = min(max(floor(local.cpu_credit_balance_period_raw / 60) * 60, 60), 86400)
+  cpu_credit_balance_evaluation_periods_raw = try(tonumber(lookup(data.aws_rds_cluster.cluster.tags, "marbot:cpu-credit-balance:evaluation-periods", var.cpu_credit_balance_evaluation_periods)), var.cpu_credit_balance_evaluation_periods)
+  cpu_credit_balance_evaluation_periods     = min(max(local.cpu_credit_balance_evaluation_periods_raw, 1), floor(86400 / local.cpu_credit_balance_period))
+
+  freeable_memory                        = lookup(data.aws_rds_cluster.cluster.tags, "marbot:freeable-memory", var.freeable_memory)
+  freeable_memory_threshold              = try(tonumber(lookup(data.aws_rds_cluster.cluster.tags, "marbot:freeable-memory:threshold", var.freeable_memory_threshold)), var.freeable_memory_threshold)
+  freeable_memory_period_raw             = try(tonumber(lookup(data.aws_rds_cluster.cluster.tags, "marbot:freeable-memory:period", var.freeable_memory_period)), var.freeable_memory_period)
+  freeable_memory_period                 = min(max(floor(local.freeable_memory_period_raw / 60) * 60, 60), 86400)
+  freeable_memory_evaluation_periods_raw = try(tonumber(lookup(data.aws_rds_cluster.cluster.tags, "marbot:freeable-memory:evaluation-periods", var.freeable_memory_evaluation_periods)), var.freeable_memory_evaluation_periods)
+  freeable_memory_evaluation_periods     = min(max(local.freeable_memory_evaluation_periods_raw, 1), floor(86400 / local.freeable_memory_period))
 }
 
 ##########################################################################
@@ -73,7 +99,7 @@ data "aws_iam_policy_document" "topic_policy" {
 
 resource "aws_sns_topic_subscription" "marbot" {
   depends_on = [aws_sns_topic_policy.marbot]
-  count      = (var.create_topic && var.enabled) ? 1 : 0
+  count      = (var.create_topic && local.enabled) ? 1 : 0
 
   topic_arn              = join("", aws_sns_topic.marbot.*.arn)
   protocol               = "https"
@@ -99,7 +125,7 @@ JSON
 
 resource "aws_cloudwatch_event_rule" "monitoring_jump_start_connection" {
   depends_on = [aws_sns_topic_subscription.marbot]
-  count      = (var.module_version_monitoring_enabled && var.enabled) ? 1 : 0
+  count      = (var.module_version_monitoring_enabled && local.enabled) ? 1 : 0
 
   name                = "marbot-rds-cluster-connection-${random_id.id8.hex}"
   description         = "Monitoring Jump Start connection. (created by marbot)"
@@ -108,7 +134,7 @@ resource "aws_cloudwatch_event_rule" "monitoring_jump_start_connection" {
 }
 
 resource "aws_cloudwatch_event_target" "monitoring_jump_start_connection" {
-  count = (var.module_version_monitoring_enabled && var.enabled) ? 1 : 0
+  count = (var.module_version_monitoring_enabled && local.enabled) ? 1 : 0
 
   rule      = join("", aws_cloudwatch_event_rule.monitoring_jump_start_connection.*.name)
   target_id = "marbot"
@@ -117,7 +143,7 @@ resource "aws_cloudwatch_event_target" "monitoring_jump_start_connection" {
 {
   "Type": "monitoring-jump-start-tf-connection",
   "Module": "rds-cluster",
-  "Version": "0.10.0",
+  "Version": "1.0.0",
   "Partition": "${data.aws_partition.current.partition}",
   "AccountId": "${data.aws_caller_identity.current.account_id}",
   "Region": "${data.aws_region.current.name}"
@@ -139,17 +165,17 @@ resource "random_id" "id8" {
 
 resource "aws_cloudwatch_metric_alarm" "cpu_utilization" {
   depends_on = [aws_sns_topic_subscription.marbot]
-  count      = (var.cpu_utilization_threshold >= 0 && var.enabled) ? 1 : 0
+  count      = (local.cpu_utilization == "static" && local.enabled) ? 1 : 0
 
   alarm_name          = "marbot-rds-cluster-cpu-utilization-${random_id.id8.hex}"
-  alarm_description   = "Average database CPU utilization over last 10 minutes too high. (created by marbot)"
+  alarm_description   = "Average database CPU utilization too high. (created by marbot)"
   namespace           = "AWS/RDS"
   metric_name         = "CPUUtilization"
   statistic           = "Average"
-  period              = 600
-  evaluation_periods  = 1
+  period              = local.cpu_utilization_period
+  evaluation_periods  = local.cpu_utilization_evaluation_periods
   comparison_operator = "GreaterThanThreshold"
-  threshold           = var.cpu_utilization_threshold
+  threshold           = local.cpu_utilization_threshold
   alarm_actions       = [local.topic_arn]
   ok_actions          = [local.topic_arn]
   dimensions = {
@@ -163,17 +189,17 @@ resource "aws_cloudwatch_metric_alarm" "cpu_utilization" {
 
 resource "aws_cloudwatch_metric_alarm" "cpu_credit_balance" {
   depends_on = [aws_sns_topic_subscription.marbot]
-  count      = (var.cpu_credit_balance_threshold >= 0 && var.burst_monitoring_enabled && var.enabled) ? 1 : 0
+  count      = (local.cpu_credit_balance == "static" && local.enabled) ? 1 : 0
 
   alarm_name          = "marbot-rds-cluster-cpu-credit-balance-${random_id.id8.hex}"
-  alarm_description   = "Average database CPU credit balance over last 10 minutes too low, expect a significant performance drop soon. (created by marbot)"
+  alarm_description   = "Average database CPU credit balance too low, expect a significant performance drop soon. (created by marbot)"
   namespace           = "AWS/RDS"
   metric_name         = "CPUCreditBalance"
   statistic           = "Average"
-  period              = 600
-  evaluation_periods  = 1
+  period              = local.cpu_credit_balance_period
+  evaluation_periods  = local.cpu_credit_balance_evaluation_periods
   comparison_operator = "LessThanThreshold"
-  threshold           = var.cpu_credit_balance_threshold
+  threshold           = local.cpu_credit_balance_threshold
   alarm_actions       = [local.topic_arn]
   ok_actions          = [local.topic_arn]
   dimensions = {
@@ -187,17 +213,17 @@ resource "aws_cloudwatch_metric_alarm" "cpu_credit_balance" {
 
 resource "aws_cloudwatch_metric_alarm" "freeable_memory" {
   depends_on = [aws_sns_topic_subscription.marbot]
-  count      = (var.freeable_memory_threshold >= 0 && var.enabled) ? 1 : 0
+  count      = (local.freeable_memory == "static" && local.enabled) ? 1 : 0
 
   alarm_name          = "marbot-rds-cluster-freeable-memory-${random_id.id8.hex}"
-  alarm_description   = "Average database freeable memory over last 10 minutes too low, performance may suffer. (created by marbot)"
+  alarm_description   = "Average database freeable memory too low, performance may suffer. (created by marbot)"
   namespace           = "AWS/RDS"
   metric_name         = "FreeableMemory"
   statistic           = "Average"
-  period              = 600
-  evaluation_periods  = 1
+  period              = local.freeable_memory_period
+  evaluation_periods  = local.freeable_memory_evaluation_periods
   comparison_operator = "LessThanThreshold"
-  threshold           = var.freeable_memory_threshold
+  threshold           = local.freeable_memory_threshold
   alarm_actions       = [local.topic_arn]
   ok_actions          = [local.topic_arn]
   dimensions = {
@@ -215,7 +241,7 @@ resource "aws_cloudwatch_metric_alarm" "freeable_memory" {
 
 resource "aws_db_event_subscription" "rds_cluster_issue" {
   depends_on = [aws_sns_topic_subscription.marbot]
-  count      = var.enabled ? 1 : 0
+  count      = local.enabled ? 1 : 0
 
   name_prefix = "marbot"
   sns_topic   = local.topic_arn

diff --git a/variables.tf b/variables.tf
@@ -1,3 +1,22 @@
+# We can not only check the var.topic_arn !="" because of the Terraform error:  The "count" value depends on resource attributes that cannot be determined until apply, so Terraform cannot predict how many instances will be created.
+variable "create_topic" {
+  type        = bool
+  description = "Create SNS topic? If set to false you must set topic_arn as well!"
+  default     = true
+}
+
+variable "topic_arn" {
+  type        = string
+  description = "Optional SNS topic ARN if create_topic := false (usually the output of the modules marbot-monitoring-basic or marbot-standalone-topic)."
+  default     = ""
+}
+
+variable "stage" {
+  type        = string
+  description = "marbot stage (never change this!)."
+  default     = "v1"
+}
+
 variable "endpoint_id" {
   type        = string
   description = "Your marbot endpoint ID (to get this value: select a channel where marbot belongs to and send a message like this: \"@marbot show me my endpoint id\")."
@@ -26,45 +45,81 @@ variable "db_cluster_identifier" {
   description = "The cluster identifier of the RDS Aurora cluster that you want to monitor."
 }
 
+
+
+variable "cpu_utilization" {
+  type        = string
+  description = "CPU utilization (static|off)."
+  default     = "static"
+}
+
 variable "cpu_utilization_threshold" {
   type        = number
-  description = "The maximum percentage of CPU utilization (set to -1 to disable)."
+  description = "The maximum percentage of CPU utilization (0-100)."
   default     = 80
 }
 
-variable "burst_monitoring_enabled" {
-  type        = bool
-  description = "Deprecated, set variable cpu_credit_balance_threshold to -1 instead"
-  default     = true
+variable "cpu_utilization_period" {
+  type        = number
+  description = "The period in seconds over which the specified statistic is applied (<= 86400 and multiple of 60)."
+  default     = 600
+}
+
+variable "cpu_utilization_evaluation_periods" {
+  type        = number
+  description = "The number of periods over which data is compared to the specified threshold (>= 1 and $period*$evaluation_periods <= 86400)."
+  default     = 1
+}
+
+
+
+
+variable "cpu_credit_balance" {
+  type        = string
+  description = "CPU burst credits for t* instances (static|off)."
+  default     = "static"
 }
 
 variable "cpu_credit_balance_threshold" {
   type        = number
-  description = "The minimum number of CPU credits available (t* instances only; set to -1 to disable)."
+  description = "The minimum number of CPU credits remaining in the burst bucket (>= 0)."
   default     = 20
 }
 
-variable "freeable_memory_threshold" {
+variable "cpu_credit_balance_period" {
   type        = number
-  description = "The minimum amount of available random access memory in Byte (set to -1 to disable)."
-  default     = 64000000 # 64 Megabyte in Byte
+  description = "The period in seconds over which the specified statistic is applied (<= 86400 and multiple of 60)."
+  default     = 600
 }
 
-# We can not only check the var.topic_arn !="" because of the Terraform error:  The "count" value depends on resource attributes that cannot be determined until apply, so Terraform cannot predict how many instances will be created.
-variable "create_topic" {
-  type        = bool
-  description = "Create SNS topic? If set to false you must set topic_arn as well!"
-  default     = true
+variable "cpu_credit_balance_evaluation_periods" {
+  type        = number
+  description = "The number of periods over which data is compared to the specified threshold (>= 1 and $period*$evaluation_periods <= 86400)."
+  default     = 1
 }
 
-variable "topic_arn" {
+
+
+variable "freeable_memory" {
   type        = string
-  description = "Optional SNS topic ARN if create_topic := false (usually the output of the modules marbot-monitoring-basic or marbot-standalone-topic)."
-  default     = ""
+  description = "Available memory (static|off)."
+  default     = "static"
 }
 
-variable "stage" {
-  type        = string
-  description = "marbot stage (never change this!)."
-  default     = "v1"
+variable "freeable_memory_threshold" {
+  type        = number
+  description = "The minimum amount of available random access memory in Byte (>= 0)."
+  default     = 64000000 # 64 Megabyte in Byte
+}
+
+variable "freeable_memory_period" {
+  type        = number
+  description = "The period in seconds over which the specified statistic is applied (<= 86400 and multiple of 60)."
+  default     = 600
+}
+
+variable "freeable_memory_evaluation_periods" {
+  type        = number
+  description = "The number of periods over which data is compared to the specified threshold (>= 1 and $period*$evaluation_periods <= 86400)."
+  default     = 1
 }