Deploy AHPA for intelligent resource prediction

Background information

The AHPA controller integrates with Managed Service for Prometheus to obtain and process historical metrics data for your application. This data forms the basis for predictive scaling decisions. Using machine learning algorithms, the AHPA controller forecasts the number of pod instances required in the next 24 hours, which is especially useful for handling workloads with cyclical patterns. By combining proactive and passive prediction strategies, the AHPA controller adjusts the number of pod instances to ensure resources are prefetched for upcoming traffic peaks. This approach enhances application responsiveness and performance while helping you manage costs and maintain service stability. For more information about AHPA, see AHPA overview.

Step 1: Install the AHPA controller

1. Log on to the ACS console. In the left navigation pane, click Clusters.

2. On the Clusters page, click the name of the target cluster. In the left navigation pane, click Add-ons.

3. On the Add-ons page, find the AHPA controller in the Others category and click Install.

Step 2: Configure Prometheus and integrate AHPA

1. Log on to the ARMS console.

2. In the left navigation pane, choose Managed Service for Prometheus > Instances.

3. At the top of the Instances page, select the region where your Prometheus instance is deployed. Click the name of the target instance, which has the same name as your ACS cluster and an Instance Type of General. In the left-side navigation pane, click Settings. In the HTTP API Address (Grafana Read Address) section, record the values of the following parameters.

   • If a token is enabled, record the access token.

   • View and record the Internal Network endpoint (Prometheus URL).

4. Set the Prometheus query URL in the ACS cluster.

   1. Create a file named application-intelligence.yaml with the following content.

      • prometheusUrl: The access endpoint of the ARMS Prometheus instance.

      • token: The access token of the Prometheus instance.

      apiVersion: v1
      kind: ConfigMap
      metadata:
        name: application-intelligence
        namespace: kube-system
      data:
        prometheusUrl: "http://cn-hangzhou-intranet.arms.aliyuncs.com:9443/api/v1/prometheus/da9d7dece901db4c9fc7f5b9c40****/158120454317****/cc6df477a982145d986e3f79c985a****/cn-hangzhou"
        token: "eyJhxxxxx"

      Note

      To view the AHPA dashboard in Managed Service for Prometheus, you must also configure the following fields in this ConfigMap:

      • prometheus_writer_url: Set this to the remote write internal endpoint.

      • prometheus_writer_ak: Set this to the AccessKey ID of your Alibaba Cloud account.

      • prometheus_writer_sk: Set this to the AccessKey secret of your Alibaba Cloud account.

   2. Run the following command to deploy application-intelligence.

      kubectl apply -f application-intelligence.yaml

5. Integrate AHPA.

   1. Log on to the ARMS console.

   2. In the left navigation pane, choose Managed Service for Prometheus > Instances.

   3. In the top navigation bar, click Integrate Other Components. In the Integration Center, search for AHPA and click the AHPA card.

   4. Integrate the AHPA component.

      1. On the ACK AHPA page, choose Select Container Service Cluster > Select Cluster, and select the cluster that you want to integrate from the drop-down list.

      2. Configure the parameters in the Configuration Information section based on the following table. Click OK.

         Parameter	Description
         Exporter name	A unique name for the AHPA exporter.
         Metrics collection interval (seconds)	The interval at which monitoring data is collected.

   5. After the Integration Status Check is complete, click Integration Management.

Step 3: Deploy a test service

The test services include fib-deployment, fib-svc, and fib-loader, a service that simulates request peaks and troughs, as well as an HPA resource to compare results with AHPA.

1. Create a file named demo.yaml with the following content.

   YAML content

   apiVersion: apps/v1
   kind: Deployment
   metadata:
     name: fib-deployment
     namespace: default
   spec:
     replicas: 1
     selector:
       matchLabels:
         app: fib-deployment
     strategy:
       rollingUpdate:
         maxSurge: 25%
         maxUnavailable: 25%
       type: RollingUpdate
     template:
       metadata:
         creationTimestamp: null
         labels:
           app: fib-deployment
       spec:
         containers:
         - image: registry.cn-huhehaote.aliyuncs.com/kubeway/knative-sample-fib-server:20200820-171837
           imagePullPolicy: IfNotPresent
           name: user-container
           ports:
           - containerPort: 8080
             name: user-port
             protocol: TCP
           resources:
             limits:
               cpu: "1"
               memory: 2000Mi
             requests:
               cpu: "1"
               memory: 2000Mi
   ---
   apiVersion: v1
   kind: Service
   metadata:
     name: fib-svc
     namespace: default
   spec:
     ports:
     - name: http
       port: 80
       protocol: TCP
       targetPort: 8080
     selector:
       app: fib-deployment
     sessionAffinity: None
     type: ClusterIP
   ---
   apiVersion: apps/v1
   kind: Deployment
   metadata:
     name: fib-loader
     namespace: default
   spec:
     progressDeadlineSeconds: 600
     replicas: 1
     revisionHistoryLimit: 10
     selector:
       matchLabels:
         app: fib-loader
     strategy:
       rollingUpdate:
         maxSurge: 25%
         maxUnavailable: 25%
       type: RollingUpdate
     template:
       metadata:
         creationTimestamp: null
         labels:
           app: fib-loader
       spec:
         containers:
         - args:
           - -c
           - |
             /ko-app/fib-loader --service-url="http://fib-svc.${NAMESPACE}?size=35&interval=0" --save-path=/tmp/fib-loader-chart.html
           command:
           - sh
           env:
           - name: NAMESPACE
             valueFrom:
               fieldRef:
                 apiVersion: v1
                 fieldPath: metadata.namespace
           image: registry.cn-huhehaote.aliyuncs.com/kubeway/knative-sample-fib-loader:20201126-110434
           imagePullPolicy: IfNotPresent
           name: loader
           ports:
           - containerPort: 8090
             name: chart
             protocol: TCP
           resources:
             limits:
               cpu: "8"
               memory: 16000Mi
             requests:
               cpu: "2"
               memory: 4000Mi
   ---
   apiVersion: autoscaling/v1
   kind: HorizontalPodAutoscaler
   metadata:
     name: fib-hpa
     namespace: default
   spec:
     maxReplicas: 50
     minReplicas: 1
     scaleTargetRef:
       apiVersion: apps/v1
       kind: Deployment
       name: fib-deployment
     targetCPUUtilizationPercentage: 50
   ---

2. Run the following command to deploy the test service.

   kubectl apply -f demo.yaml

Step 4: Create an AHPA resource

Submit an AdvancedHorizontalPodAutoscaler resource to configure the scaling policy.

1. Create a file named ahpa-demo.yaml with the following content.

   apiVersion: autoscaling.alibabacloud.com/v1beta1
   kind: AdvancedHorizontalPodAutoscaler
   metadata:
     name: ahpa-demo
   spec:
     scaleStrategy: observer
     metrics:
     - type: Resource
       resource:
         name: cpu
         target:
           type: Utilization
           averageUtilization: 40
     scaleTargetRef:
       apiVersion: apps/v1
       kind: Deployment
       name: fib-deployment 
     maxReplicas: 100
     minReplicas: 2
     stabilizationWindowSeconds: 300
     prediction:
       quantile: 95
       scaleUpForward: 180
     instanceBounds:
     - startTime: "2021-12-16 00:00:00"
       endTime: "2031-12-16 00:00:00"
       bounds:
       - cron: "* 0-8 ? * MON-FRI"
         maxReplicas: 15
         minReplicas: 4
       - cron: "* 9-15 ? * MON-FRI"
         maxReplicas: 15
         minReplicas: 10
       - cron: "* 16-23 ? * MON-FRI"
         maxReplicas: 20
         minReplicas: 15

   The following table describes some of the parameters.

   Parameter	Required	Description
   scaleTargetRef	Yes	Specifies the target Deployment.
   metrics	Yes	Configures the metrics for scaling. Supported metrics include CPU, GPU, memory, QPS, and RT.
   target	Yes	The target threshold. For example, averageUtilization: 40 means the target CPU utilization is 40%.
   scaleStrategy	No	Specifies the scaling mode. The default value is observer. auto: AHPA manages scaling. observer: Observes but does not perform actual scaling actions. You can use this mode to check whether AHPA works as expected. scalingUpOnly: Performs scale-out operations only. proactive: Enables only proactive prediction. reactive: Enables only passive prediction.
   maxReplicas	Yes	The maximum number of replicas for scale-out.
   minReplicas	Yes	The minimum number of replicas for scale-in.
   stabilizationWindowSeconds	No	The cooldown period for scale-in. Default value: 300 seconds.
   prediction.quantile	Yes	The prediction quantile. A higher value is more conservative because it increases the probability that the actual metric value stays below the target. Valid values: 0 to 100. Default value: 99. Recommended range: 90 to 99.
   prediction. scaleUpForward	Yes	The time required for a new pod to become ready (the cold start time).
   instanceBounds	No	The boundaries for the number of replicas within specific time ranges. startTime: The start time. endTime: The end time.
   instanceBounds.bounds.cron	No	A cron expression that specifies a schedule. A cron expression consists of five space-separated fields. For example, - cron: "* 0-8 ? * MON-FRI" specifies that the task runs from 00:00 to 08:59 every Monday to Friday.

   The following table describes the fields of a cron expression. For more information, see Cron expressions.

   Field	Required	Allowed values	Allowed special characters
   Minutes	Yes	0-59	* / , -
   Hours	Yes	0-23	* / , -
   Day of month	Yes	1-31	* / , – ?
   Month	Yes	1-12 or JAN-DEC	* / , -
   Day of week	No	0-6 or SUN-SAT	* / , – ?

   Note

   • The values of the Month and Day of week fields are not case-sensitive. For example, SUN, Sun, and sun have the same effect.

   • If the Day of week field is not configured, the default value is *.

   • Special characters:

     • *: specifies all possible values.

     • /: specifies an increment.

     • ,: lists values.

     • -: specifies a range.

     • ?: specifies no specific value.

2. Run the following command to create the AHPA scaling policy.

   kubectl apply -f ahpa-demo.yaml

Step 5: View the prediction results

On the Integration Management page, navigate to the Container Service tab and click your cluster name. Select ACK AHPA from the Component Type drop-down list. Finally, go to the Dashboards tab and click ahpa-dashboard to view the monitoring data.

This topic provides an example where the scaling mode is configured as observer mode (observer mode). This example compares the results with an HPA policy, which serves as a reference for the actual resources required by the application, to check whether the AHPA prediction results are as expected.