Hello, and welcome. Orlando Gentil here. Today, I will demonstrate how to configure autoscaling on OCI. Very first step to configure autoscaling is to create the instance configuration based on an existing instance. Let's take a look how we can do that. From our previous demo, we have already our web server running, so we're going to go to the navigation menu, select Compute, Instances, and in the compartment that we used before, the sandbox compartment, we're going to see our webserver01. Click on the webserver01, and then under more actions, you will have Create Instance Configuration. Instance-config, I'm going to rename it to instance-config demo. The compartment where it will exist, and now you click on "Create Instance configuration". As we create the pool, we will take advantage and also configure it. From within the instance configuration details, we're going to have a button to create the instance pool. Let's click on it. Let's name our pool, I'm going to call it demo, to be consistent. On the instance configuration details, you have just a summary of all the information that is present on the instance configuration. Let's set the number of instances for one. When the pool is created, we are going to have one instance. Let's go to the next page. In the pool placement screen, you have the availability domains where the instances that will be created by the pool where they will be placed, which of availability domains and we think the availability domains, you have to select what's going to be the VCN, that they will be part of and the subnet where the phoenix will be placed. As we have more than one availability domain in [inaudible], let's go ahead and select the second availability domain. We want to change the default domains. The VNIC is going to be on the same VCN. Let's go with the same subnet. We didn't create a load balancer to attach this pole to there, so if we had a load balancer, here is where we would select, but in this case we don't have. Let's skip that part. Let's review the information that we just added in the previous screens. Let's create the pool. As it starts provisioning the pool, as we set the initial number to be one, the very first operation that we see here in the work requests is to create instances in pool. It starts to populate the pool with that first instance. Now that the pool finished provisioning and it's running, it's time for our final step. Let's create an autoscaling configuration. To create the autoscaling configuration, we can just come from the pool details, and "More Actions", "Create Autoscaling Configuration". Let's keep domain consistent. Let's change it to demo. Let's move to the next. We're going to pick metric-based. If the threshold that we said is exceeded or CPU utilization or memory, it will trigger the creation of new instances. If we go with schedule-based it's based on time. But that's not what you're going to use today. Instance pool configuration is just a summary of what we already have on that pool. Let us start configuring our policy. Let's keep it consistent. The cooldown seconds. The interval between scaling that we have, we're going to pick a CPU utilization. We will define what's going to be this scale-out rule. We're going to go with greater than. The threshold that we're going to set is if it goes beyond 80 percent of utilization, it's going to add one instance. We have set the scale-in rule. Whenever it goes below 50, it's going to remove one instance. In this Scaling limits, we set the number of instances that we have. We're going to start setting the minimum number of instances. As the initial number is one, we can't go with less than one. With the maximum, we're going to set how many instances we want to have on the pool following the Scale-out rule. We're going to go with Next. We have a summary of our options, and we're going with Create. It's provisioned. Now, it's time to push one of the instances to exceed the threshold and see what's going to happen. Let's go back to the Instance Pool list, let's select our demo pool, and from the Resources, we can go to Attached Instances, and we can see the instances that are currently running. As we set the initial size to one, we already see this instance here. This is the instance that we're going to stress. To do that, I will have to connect to this instance using SSH, and for that I'm going to use Cloud Shell, ssh opc@. Remember from the previous demo that we already have the SSH key set and configured on the default directory, so I don't have to specify anything here. It's the first time that we're connecting to that instance. Let's make sense. Let's clear first here. I have to Install the stress package. The stress package can be found on the developer's repository that we have on OCI. That repo is not enabled by default, so let's enable that repo by running sudo yum-config-manager --enable ol7_develper_EPEL; is the name of the repo. After that, I will Install stress package. Confirm is yes. Now that the stress package is installed, let's use it to force the CPU utilization stress. Let's monitor the pool and see what's going to happen. Pool, it's demo. For the metrics here, as soon as it starts to spike, we'll be able to monitor here, and the status is going to change here from Running to Scaling. Let's keep an eye on that. After 4-5 minutes , it starts Scaling. Let's take a look here at the Attached Instances and a new instance has been created. Let's wait until it finishes. Well, it finished provisioning. The second instance is showing as Running, and the status of the pool changes to Running as well. Let's stop it here, the stress program. Let's clear the screen, and let's go back on monitoring and see when the load goes down, what's going to happen here. As it will take five minutes to do the cool-down, I will pause and get back on the video as soon as it changes the status. After the cool-down period is over, now it's time to scale down. If we look at the list of the instances, where is it? Here. Instance Pool Details, it's just Scaling down. Attached Instances. Let's keep an eye here. As it's Scaling down, the second instance now it's turning to Terminating. The pool is getting back to the original size. The instance is terminated, the pool is back to the initial one instance. This concludes our demo on how to set up autoscaling. We demonstrated it working step-by-step, and forcing one of the instances to go beyond the threshold, and kicking to create another instance.
