Deploying the Horizontal Pod Autoscaler

12.4 Deploying the Horizontal Pod Autoscaler

The steps below show how to configure and run the Horizontal Pod Autoscaler (HPA) to scale Oracle Unified Directory (OUD), based on the CPU or memory utilization resource metrics.

Assuming the example OUD configuration in Creating Oracle Unified Directory Instances, three OUD servers are started by default (oud-ds-rs-0, oud-ds-rs-1, oud-ds-rs-2).

In the following example an HPA resource is created, targeted at the statefulset oud-ds-rs. This resource will autoscale OUD servers from a minimum of 3 OUD servers up to 5 OUD servers. Scaling up will occur when the average CPU is consistently over 70%. Scaling down will occur when the average CPU is consistently below 70%.

Find the statefulset used by OUD:
```
kubectl get statefulset -n <namespace>
```
For example:
```
kubectl get statefulset -n oudns
```
The output will look similar to the following:
```
NAME        READY   AGE
oud-ds-rs   3/3     23h
```

Navigate to the $WORKDIR/kubernetes/hpa and create an autoscalehpa.yaml file that contains the following:

#
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: oud-sts-hpa
  namespace: oudns
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: StatefulSet
    name: oud-ds-rs #statefulset name of oud
  behavior:
    scaleDown:
      stabilizationWindowSeconds: 60
    scaleUp:
      stabilizationWindowSeconds: 60
  minReplicas: 3
  maxReplicas: 5
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 70

where:

oud-ds-rs is the stateful set returned earlier
minReplicas should match the number of OUD servers started by default.
maxReplicas should be set to the maximum amount of OUD servers that can be started by HPA.

Note:

For setting HPA based on Memory Metrics, update the metrics block with the following content. Please note, Oracle recommends using only CPU or Memory, not both:

metrics:
- type: Resource
  resource:
    name: memory
    target:
      type: Utilization
      averageUtilization: 70

Run the following command to create the autoscaler:
```
kubectl apply -f autoscalehpa.yaml
```
The output will look similar to the following:
```
horizontalpodautoscaler.autoscaling/oud-sts-hpa created
```

Verify the status of the autoscaler by running the following:

kubectl get hpa -n oudns

The output will look similar to the following:

NAME          REFERENCE               TARGETS   MINPODS   MAXPODS   REPLICAS   AGE
oud-sts-hpa   StatefulSet/oud-ds-rs   5%/70%    3         5         3          33s