Cloudwatch alarms insufficient data

[chelomontilla]

I've created a redis without enabling the automatic_failover, only one node. The problem is that alarms at cloudwatch shows insufficient data.
The configured dimension value for CacheClusterId in the alarm is "myredis" and if I see in cloudwatch the current metrics having data are in dimension value "myredis-001".

[mukill]

were you able to resolve this?

[chelomontilla]

No, still not

[varunpalekar]

+1

[chrischildresssg]

The module itself doesn't include necessary code to handle insufficient data. Both of the "aws_cloudwatch_metric_alarm" resources need to have the "treat_missing_data" argument presented, but do not.

For instance, the "cache_memory" alarm should look like this, with the "treat_missing_data" argument near the bottom:

resource "aws_cloudwatch_metric_alarm" "cache_memory" {
count = var.enabled ? 1 : 0
alarm_name = "${module.label.id}-freeable-memory"
alarm_description = "Redis cluster freeable memory"
comparison_operator = "LessThanThreshold"
evaluation_periods = "1"
metric_name = "FreeableMemory"
namespace = "AWS/ElastiCache"
period = "60"
statistic = "Average"

threshold = var.alarm_memory_threshold_bytes

dimensions = {
CacheClusterId = module.label.id
}

alarm_actions = var.alarm_actions
ok_actions = var.ok_actions

treat_missing_data = var.treat_missing_data

depends_on = [aws_elasticache_replication_group.default]
}

[varunpalekar]

I found the main problem, it's due to creating more than one node in cluster mode, and for single also we need to add node number.
So may be a count with alarm and counting mechanism with module.label.id would solve the problem

dimensions = {
    CacheClusterId = module.label.id
}

I will also create an MR for this.