注册 ECS fargate 实例时 NLB 健康检查随机失败

Question

I'm seeing a random failure of NLB health checks when registering the ECS fargate instances, healths checks get passed after a couple of failures. I do a wide open SG that's attached to fargate Instances. Did anyone had a similar behaviour while registering the tasks under NLB targegroups?

Answer 1

You application can take some time before it starts responding to the health checks from the ELB.

When you create a ECS service, there is an option called healthCheckGracePeriodSeconds . It governs how many seconds ECS scheduler will ignore health checks information from the ELB. This option is only available if you use ELB.

So I recomend you to play with it and pick a suitable time frame for you.