Skip to content

Auto-generated

This page is generated by terraform-docs at build time from the infra/ repo. Do not edit manually.

Monitoring — Application Load Balancer

CloudWatch alarms for an ALB: - HTTPCode_ELB_5XX_Count — the ALB itself returning 5xx - HTTPCode_Target_5XX_Count — sustained app-side 5xx from targets - TargetResponseTime — average request latency, loose threshold - UnHealthyHostCount — one alarm per target group (via for_each)

Requirements

No requirements.

Providers

Name Version
aws n/a

Modules

No modules.

Resources

Name Type
aws_cloudwatch_metric_alarm.elb_5xx resource
aws_cloudwatch_metric_alarm.target_5xx resource
aws_cloudwatch_metric_alarm.target_response_time resource
aws_cloudwatch_metric_alarm.unhealthy_hosts resource

Inputs

Name Description Type Default Required
alarm_actions List of ARNs to notify on alarm transitions. Typically the PagerDuty SNS topic ARN. list(string) n/a yes
alb_arn_suffix ALB ARN suffix (the LoadBalancer CloudWatch dimension), e.g. app//. string n/a yes
alb_name ALB name, used as the alarm-name prefix. string n/a yes
base_tags Base tags applied to all alarms. map(string) {} no
elb_5xx_evaluation_periods Consecutive periods that must breach for the ELB-5xx alarm to fire. number 1 no
elb_5xx_threshold Trigger when HTTPCode_ELB_5XX_Count exceeds this in a single period. number 50 no
latency_evaluation_periods Consecutive periods that must breach for the latency alarm to fire. number 3 no
latency_seconds_threshold Trigger when TargetResponseTime average exceeds this many seconds. number 2 no
period Period in seconds for each evaluation. number 300 no
target_5xx_evaluation_periods Consecutive periods that must breach for the target-5xx alarm to fire. number 2 no
target_5xx_threshold Trigger when HTTPCode_Target_5XX_Count exceeds this in each period. number 100 no
target_groups Map of label -> target group ARN suffix. One UnHealthyHostCount alarm per entry. Empty map disables those alarms. map(string) {} no
unhealthy_evaluation_periods Consecutive periods that must breach for the unhealthy-host alarm to fire. number 2 no
unhealthy_threshold Trigger when UnHealthyHostCount is at or above this. 1 by default — even one unhealthy host is worth knowing about. number 1 no

Outputs

No outputs.