Enable request retry feature using AppQoE for NetScaler Ingress Controller
When a NetScaler appliance receives an HTTP request and forwards it to a back-end server, sometimes there may be connection failures with the back-end server. You can configure the request-retry feature on NetScaler to forward the request to the next available server, instead of sending the reset to the client. Hence, the client saves round trip time when NetScaler initiates the same request to the next available service. For more information request retry feature, see the NetScaler documentation
Now, you can configure request retry on NetScaler with NetScaler Ingress Controller. Custom Resource Definitions (CRDs) are the primary way of configuring policies in cloud native deployments. Using the AppQoE CRD provided by NetScaler, you can configure request-retry policies on NetScaler with the NetScaler Ingress Controller. The AppQoE CRD enables communication between the NetScaler Ingress Controller and NetScaler for enforcing AppQoE policies.
AppQoE CRD definition
The AppQoE CRD is available in the NetScaler Ingress Controller GitHub repo at: appqoe-crd.yaml. The AppQoE CRD provides attributes for the various options that are required to define the AppQoE policy on NetScaler.
The following are the attributes provided in the AppQoE CRD:
||Specifies the list of Kubernetes services to which you want to apply the AppQoE policies.|
||Specifies whether to set retry on connection Reset or Not|
||Specifies the time in milliseconds for retry|
||Specifies the number of retries|
||Specifies the expression for evaluating traffic.|
||Specifies the bind point for binding the AppQoE policy.|
Deploy the AppQoE CRD
Perform the following to deploy the AppQoE CRD:
Download the AppQoE CRD.
Deploy the AppQoE CRD using the following command:
kubectl create -f appqoe-crd.yaml
After you have deployed the AppQoE CRD provided by NetScaler in the Kubernetes cluster, you can define the AppQoE policy configuration in a
.yaml file. In the
.yaml file, use
appqoepolicy in the kind field and in the
spec section add the AppQoE CRD attributes based on your requirement for the policy configuration.
The following YAML file applies the AppQoE policy to the services listed in the servicenames field. You must configure the AppQoE action to retry on timeout and define the number of retry attempts.
apiVersion: citrix.com/v1 kind: appqoepolicy metadata: name: targeturlappqoe spec: appqoe-policies: - servicenames: - apache appqoe-policy: operation-retry: onReset: 'YES' onTimeout: 33 number-of-retries: 2 appqoe-criteria: 'HTTP.REQ.HEADER("User-Agent").CONTAINS("Android")' direction: REQUEST
After you have defined the policy configuration, deploy the
.yaml file using the following commands:
$ kubectl create -f appqoe-example.yaml