Added HealthyHostCount alarm

[tovbinm]

No description provided.

[lorenzoaiello]

Hey @tovbinm - Thanks for the PR.

Just curious - what's the use case for having an alarm fire when the healthy host count is greater than zero? I could see one in case the healthy host count equals zero.

[tovbinm]

Unfortunately UnHealthyHostCount metrics does not capture the case where a host (e.g. ECS Container) fails to start, since it never reaches Unhealthy state. Instead the container just bounces from Initial to Draining - https://docs.aws.amazon.com/elasticloadbalancing/latest/application/target-group-health-checks.html#target-health-states

Using HealthyHostCount lets you capture this.

[lorenzoaiello]

Gotcha, but in that scenario wouldn't you want to use LessThanOrEqualToThreshold 0 instead of GreaterThanThreshold 0?

[lorenzoaiello]

lgtm - thanks for the fix.

[tovbinm]

Thanks @lorenzoaiello !

It's also worth mentioning in the README that target_group_id and load_balancer_id should be set to use tg.arn_suffix and lb.arn_suffix respectively - e.g [1] and [2]