Hi,
I have a question that is bothering me. I am using Anislbe to do the rolling deploy on Amazon. Have set the health check to be ELB, set up all the intervals, number of checks or the instance to be healthy and unhealthy, set the health check grace period. But it seems like health check grace period is ignored, even if I put it to 10min, after 2 minutes when instance becomes in service, ASG is terminating the old instance and starts creating another one. Not very good with pyhton, but in the module found this comment:
if the health_check_type is ELB, we want to query the ELBs directly for instance
status as to avoid health_check_grace period that is awarded to ASG instances
Does that mean, that when using rolling deploy with ansible, Health check grace period is ignored, and ASG just checks what elasticloadbalancer says about instance health.
So basically as it takes about 2 minutes from when the instance is started to be fully ready, i need to create health checks so instance is not shown as out of service during those 2 minutes?
So basically, just wondering, if grace period is totally skipped, as, if i understood it correctly, ASG should wait for that period until it checks what is the instance status? But when i put that period for some greater value, it does not wait that long.
Thanks
Uros