During the coverage of EC2 Auto-Scaling, you refer to two cooldown periods, one for health checks which allows a new instance to fully launch before being required to respond to health checks and another for how long scaling should wait before making a scaling decision after the launch of the instance. Both of these are referred to in the videos as "cooldowns", but isn’t the 300 seconds wait before scaling decisions are made called a "Warm-Up"? Just wondering if referring to both as cooldowns is acceptable or if the default 300 seconds pause for scaling decisions after a scaling event would be known as a "Warm-Up" period on the exam. Thanks!
I think I do eventually use distinct terms in the lecture, but Warm-Up is more associated with the time it takes a new instance to get stable before we expect it to start responding to health checks. Cooldown period is the "timeout" before responding to new scaling events. You can read in more detail here: https://docs.aws.amazon.com/autoscaling/ec2/userguide/as-scaling-simple-step.html
I got a bit confused. From what I understood is that Cooldown period has nothing to do with health check grace-period. These are two separate things. The cooldown period helps you prevent your Auto Scaling group from launching or terminating additional instances before the effects of previous activities are visible.
While the health check grace-period is the amount of time until Amazon EC2 Auto Scaling checks the health of instances after they are put into service. The intention of this setting is to prevent Amazon EC2 Auto Scaling from marking instances as unhealthy and terminating them before they have time to come up.
But the health check grace-period and the warm-up, are basically the same thing, since AWS defines the warm-up as the number of seconds that it takes for a newly launched instance to warm up. Until its specified warm-up time has expired, an instance is not counted toward the aggregated metrics of the Auto Scaling group.
Thank you for your insights
Thanks for the explanation, Scott. I jumped into the console and saw where I made my mistake.. For a Step policy they refer to the 300 seconds as a warm-up; after each step the instance doesn’t contribute to scaling metrics for a default of 300 seconds. A Simple scaling policy refers to this period as a "cooldown", but seems to accomplish the same goal as the warm-up of a step function; no scaling actions are allowed for 300 seconds by default after a scale-out/in.. THEN, on top of that you have the health check grace-period; which luckily also default to 300 seconds (easier to remember).. lol I was tripping up on there being different names between the behavior in a step vs. a simple scaling policy.. one being a warm-up and the other a cooldown. My apologies if you covered it in the video, it’s possible I missed it, Sir. Thanks again for your time and efforts to the course, much appreciated!
I suppose the real delta between a "warm-up" and the "cooldown" is that with a cooldown there cannot be anymore scaling until the instance is up for X seconds, BUT with the warm-up, it’s possible to have scaling events. The new instance simply wouldn’t count towards the metrics that affected a scaling decision. Right?
Same doubt here. While the instances are being warmed up. will auto-scaling trigger new events to increase the capacity? If so, If a instance has a long period to warm-up it will increase in the wrong manner.