Verifying HPA

16.5 Verifying HPA

To verify the Horizontal Pod Autoscaler (HPA) works, perform the following steps:

Check the current status of the Oracle Access Management (OAM) servers:

kubectl get pods -n oamns

The output will look similar to the following:

NAME                                                     READY   STATUS      RESTARTS        AGE
accessdomain-adminserver                                 0/1     Running     0               141m
accessdomain-oam-policy-mgr1                             0/1     Running     0               138m
accessdomain-oam-server1                                 1/1     Running     0               138m

In the above example only accessdomain-oam-server1 is running.

To test HPA can scale up the WebLogic cluster accessdomain-oam-cluster, run the following commands:
```
kubectl exec --stdin --tty accessdomain-oam-server1 -n oamns -- /bin/bash
```
This will take you inside a bash shell inside the oam_server1 pod:
```
[oracle@accessdomain-oam-server1 oracle]$
```
Inside the bash shell, run the following command to increase the load on the CPU:
```
[oracle@accessdomain-oam-server1 oracle]$ dd if=/dev/zero of=/dev/null
```
This command will continue to run in the foreground.
In a command window outside the bash shell, run the following command to view the current CPU usage:
```
kubectl get hpa -n oamns
```
The output will look similar to the following:
```
NAME                           REFERENCE                          TARGETS   MINPODS   MAXPODS   REPLICAS   AGE
accessdomain-oam-cluster-hpa   Cluster/accessdomain-oam-cluster   470%/70%    1         5         1        21s
```
In the above example the CPU has increased to 470%. As this is above the 70% limit, the autoscaler increases the replicas on the Cluster resource, and the operator responds by starting additional cluster members.

Run the following to see if any more OAM Managed Servers are started:

kubectl get pods -n oamns

The output will look similar to the following:

NAME                                                     READY   STATUS      RESTARTS        AGE
accessdomain-adminserver                                 0/1     Running                     143m
accessdomain-oam-policy-mgr1                             0/1     Running     0               140m
accessdomain-oam-server1                                 1/1     Running     0               140m
accessdomain-oam-server2                                 1/1     Running     0               3m20s
accessdomain-oam-server3                                 1/1     Running     0               3m20s
accessdomain-oam-server4                                 1/1     Running     0               3m19s
accessdomain-oam-server5                                 1/1     Running     0               3m5s

In the example above four more OAM managed servers have been started (oam-server2 - oam-server5).

Note:

It may take some time for the server to appear and start. Once the servers are at READY status of 1/1, the servers are started.

To stop the load on the CPU, in both bash shells, issue a Control C, and then exit the bash shell:
```
[oracle@accessdomain-oam-server1 oracle]$ dd if=/dev/zero of=/dev/null
^C
[oracle@accessdomain-oam-server1 oracle]$ exit
```

Run the following command to view the current CPU usage:

kubectl get hpa -n oamns

The output will look similar to the following:

NAME                           REFERENCE                          TARGETS   MINPODS   MAXPODS   REPLICAS   AGE
accessdomain-oam-cluster-hpa   Cluster/accessdomain-oam-cluster   19%/70%   1         5         5          19m

In the above example CPU has dropped to 19%. As this is below the 70% threshold, you should see the autoscaler scale down the servers:

kubectl get pods -n oamns

The output will look similar to the following:

NAME                                                     READY   STATUS        RESTARTS        AGE
accessdomain-adminserver                                 1/1     Running       0               152m
accessdomain-oam-policy-mgr1                             1/1     Running       0               149m
accessdomain-oam-server1                                 1/1     Running       0               149m
accessdomain-oam-server2                                 1/1     Running       0               14m
accessdomain-oam-server3                                 0/1     Terminating   0               14m

Eventually, all the servers except oam-server1 will disappear:

NAME                                                     READY   STATUS      RESTARTS       AGE
accessdomain-adminserver                                 1/1     Running     0              154m
accessdomain-oam-policy-mgr1                             1/1     Running     0              151m
accessdomain-oam-server1                                 1/1     Running     0              151m