MON-112 add alb monitors

This commit is contained in:
Quentin Manfroi 2018-01-26 19:07:22 +01:00
parent 37d83aa0f2
commit 3cb93f967c
3 changed files with 304 additions and 0 deletions

53
cloud/aws/alb/README.md Normal file
View File

@ -0,0 +1,53 @@
AWS ALB DataDog monitors
==========================================
How to use this module
----------------------
```
module "datadog-monitors-aws-alb" {
source = "git::ssh://git@bitbucket.org/morea/terraform.feature.datadog.git//cloud/aws/alb?ref={revision}"
environment = "${var.environment}"
message = "${module.datadog-message-alerting.alerting-message}"
}
```
Purpose
-------
Creates DataDog monitors with the following checks :
* ALB no healthy hosts
* ALB latency too high
* ALB http code 5xx percent to high
* ALB http code 4xx percent to high
* ALB target http code 5xx percent to high
* ALB target http code 4xx percent to high
Inputs
------
| Name | Description | Type | Default | Required |
|------|-------------|:----:|:-----:|:-----:|
| delay | Delay in seconds for the metric evaluation | string | `900` | no |
| environment | Architecture environment | string | - | yes |
| filter_tags | Tags used for custom filtering | string | `*` | no |
| httpcode_elb_4xx_threshold_critical | loadbalancer 4xx critical threshold in percentage | string | `80` | no |
| httpcode_elb_4xx_threshold_warning | loadbalancer 4xx warning threshold in percentage | string | `60` | no |
| httpcode_elb_5xx_threshold_critical | loadbalancer 5xx critical threshold in percentage | string | `80` | no |
| httpcode_elb_5xx_threshold_warning | loadbalancer 5xx warning threshold in percentage | string | `60` | no |
| httpcode_target_4xx_threshold_critical | target 4xx critical threshold in percentage | string | `80` | no |
| httpcode_target_4xx_threshold_warning | target 4xx warning threshold in percentage | string | `60` | no |
| httpcode_target_5xx_threshold_critical | target 5xx critical threshold in percentage | string | `80` | no |
| httpcode_target_5xx_threshold_warning | target 5xx warning threshold in percentage | string | `60` | no |
| message | Message sent when a monitor is triggered | string | - | yes |
| latency_threshold_critical | latency critical threshold in milliseconds | string | `1000` | no |
| latency_threshold_warning | latency warning threshold in milliseconds | string | `500` | no |
Related documentation
---------------------
DataDog blog: [https://www.datadoghq.com/blog/monitor-application-load-balancer/](https://www.datadoghq.com/blog/monitor-application-load-balancer/)
AWS ALB metrics documentation: [https://docs.aws.amazon.com/elasticloadbalancing/latest/application/load-balancer-cloudwatch-metrics.html](https://docs.aws.amazon.com/elasticloadbalancing/latest/application/load-balancer-cloudwatch-metrics.html)

78
cloud/aws/alb/inputs.tf Normal file
View File

@ -0,0 +1,78 @@
# Datadog global variables
variable "environment" {
description = "Architecture environment"
type = "string"
}
variable "filter_tags_use_defaults" {
description = "Use default filter tags convention"
default = "true"
}
variable "filter_tags_custom" {
description = "Tags used for custom filtering when filter_tags_use_defaults is false"
default = "*"
}
variable "message" {
description = "Message sent when a monitor is triggered"
}
variable "delay" {
description = "Delay in seconds for the metric evaluation"
default = 900
}
# Datadog monitors variables
variable "latency_threshold_critical" {
default = 1000
description = "latency critical threshold in milliseconds"
}
variable "latency_threshold_warning" {
default = 500
description = "latency warning threshold in milliseconds"
}
variable "httpcode_elb_4xx_threshold_critical" {
default = 80
description = "loadbalancer 4xx critical threshold in percentage"
}
variable "httpcode_elb_4xx_threshold_warning" {
default = 60
description = "loadbalancer 4xx warning threshold in percentage"
}
variable "httpcode_target_4xx_threshold_critical" {
default = 80
description = "target 4xx critical threshold in percentage"
}
variable "httpcode_target_4xx_threshold_warning" {
default = 60
description = "target 4xx warning threshold in percentage"
}
variable "httpcode_elb_5xx_threshold_critical" {
default = 80
description = "loadbalancer 5xxcritical threshold in percentage"
}
variable "httpcode_elb_5xx_threshold_warning" {
default = 60
description = "loadbalancer 5xx warning threshold in percentage"
}
variable "httpcode_target_5xx_threshold_critical" {
default = 80
description = "target 5xx critical threshold in percentage"
}
variable "httpcode_target_5xx_threshold_warning" {
default = 60
description = "target 5xx warning threshold in percentage"
}

View File

@ -0,0 +1,173 @@
data "template_file" "filter" {
template = "$${filter}"
vars {
filter = "${var.filter_tags_use_defaults == "true" ? format("dd_monitoring:enabled,dd_aws_alb:enabled,env:%s", var.environment) : "${var.filter_tags_custom}"}"
}
}
resource "datadog_monitor" "ALB_no_healthy_instances" {
name = "[${var.environment}] ALB no healthy instances"
type = "metric alert"
message = "${var.message}"
query = <<EOF
min(last_1m): (
sum:aws.applicationelb.healthy_host_count{${var.filter_tags}} by {region,loadbalancer}
) <= 0
EOF
evaluation_delay = "${var.delay}"
new_host_delay = "${var.delay}"
thresholds {
critical = 0
}
notify_no_data = true # Will notify when no data is received
renotify_interval = 0
require_full_window = false
timeout_h = 0
include_tags = true
tags = ["env:${var.environment}", "resource:alb", "team:aws", "provider:aws"]
}
resource "datadog_monitor" "ALB_latency" {
name = "[${var.environment}] ALB latency > ${var.latency_threshold_critical} ms"
type = "metric alert"
message = "${var.message}"
query = <<EOF
avg(last_5m): (
avg:aws.applicationelb.target_response_time.average{${var.filter_tags}} by {region,loadbalancer}
) > ${var.latency_threshold_critical}
EOF
evaluation_delay = "${var.delay}"
new_host_delay = "${var.delay}"
thresholds {
critical = ${var.latency_threshold_critical}
warning = ${var.latency_threshold_warning}
}
notify_no_data = true # Will notify when no data is received
renotify_interval = 0
require_full_window = false
timeout_h = 0
include_tags = true
tags = ["env:${var.environment}", "resource:alb", "team:aws", "provider:aws"]
}
resource "datadog_monitor" "ALB_httpcode_elb_5xx" {
name = "[${var.environment}] ALB HTTP code 5xx > ${var.httpcode_elb_5xx_threshold_critical} %"
type = "metric alert"
message = "${var.message}"
query = <<EOF
sum(last_5m): (
avg:aws.applicationelb.httpcode_elb_5xx{${var.filter_tags}} by {region,loadbalancer}.as_count() /
avg:aws.applicationelb.request_count{${var.filter_tags}} by {region,loadbalancer}.as_count() * 100
) > ${var.httpcode_elb_5xx_threshold_critical}
EOF
evaluation_delay = "${var.delay}"
new_host_delay = "${var.delay}"
thresholds {
critical = ${var.httpcode_elb_5xx_threshold_critical}
warning = ${var.httpcode_elb_5xx_threshold_warning}
}
notify_no_data = false # Will notify when no data is received
renotify_interval = 0
require_full_window = false
timeout_h = 1
include_tags = true
tags = ["env:${var.environment}", "resource:alb", "team:aws", "provider:aws"]
}
resource "datadog_monitor" "ALB_httpcode_elb_4xx" {
name = "[${var.environment}] ALB HTTP code 4xx > ${var.httpcode_elb_4xx_threshold_critical} %"
type = "metric alert"
message = "${var.message}"
query = <<EOF
sum(last_5m): (
avg:aws.applicationelb.httpcode_elb_4xx{${var.filter_tags}} by {region,loadbalancer}.as_count() /
avg:aws.applicationelb.request_count{${var.filter_tags}} by {region,loadbalancer}.as_count() * 100
) > ${var.httpcode_elb_4xx_threshold_critical}
EOF
evaluation_delay = "${var.delay}"
new_host_delay = "${var.delay}"
thresholds {
critical = ${var.httpcode_elb_4xx_threshold_critical}
warning = ${var.httpcode_elb_4xx_threshold_warning}
}
notify_no_data = false # Will notify when no data is received
renotify_interval = 0
require_full_window = false
timeout_h = 1
include_tags = true
tags = ["env:${var.environment}", "resource:alb", "team:aws", "provider:aws"]
}
resource "datadog_monitor" "ALB_httpcode_target_5xx" {
name = "[${var.environment}] ALB target HTTP code 5xx > ${var.httpcode_target_5xx_threshold_critical} %"
type = "metric alert"
message = "${var.message}"
query = <<EOF
sum(last_5m): (
avg:aws.applicationelb.httpcode_target_5xx{${var.filter_tags}} by {region,loadbalancer}.as_count() /
avg:aws.applicationelb.request_count{${var.filter_tags}} by {region,loadbalancer}.as_count() * 100
) > ${var.httpcode_target_5xx_threshold_critical}
EOF
evaluation_delay = "${var.delay}"
new_host_delay = "${var.delay}"
thresholds {
critical = ${var.httpcode_target_5xx_threshold_critical}
warning = ${var.httpcode_target_5xx_threshold_warning}
}
notify_no_data = false # Will notify when no data is received
renotify_interval = 0
require_full_window = false
timeout_h = 1
include_tags = true
tags = ["env:${var.environment}", "resource:alb", "team:aws", "provider:aws"]
}
resource "datadog_monitor" "ALB_httpcode_target_4xx" {
name = "[${var.environment}] ALB target HTTP code 4xx > ${var.httpcode_target_4xx_threshold_critical} %"
type = "metric alert"
message = "${var.message}"
query = <<EOF
sum(last_5m): (
avg:aws.applicationelb.httpcode_target_4xx{${var.filter_tags}} by {region,loadbalancer}.as_count() /
avg:aws.applicationelb.request_count{${var.filter_tags}} by {region,loadbalancer}.as_count() * 100
) > ${var.httpcode_target_4xx_threshold_critical}
EOF
evaluation_delay = "${var.delay}"
new_host_delay = "${var.delay}"
thresholds {
critical = ${var.httpcode_target_4xx_threshold_critical}
warning = ${var.httpcode_target_4xx_threshold_warning}
}
notify_no_data = false # Will notify when no data is received
renotify_interval = 0
require_full_window = false
timeout_h = 1
include_tags = true
tags = ["env:${var.environment}", "resource:alb", "team:aws", "provider:aws"]
}