This document explains how to create your custom values file for the Scalar Envoy chart. If you want to know the details of the parameters, please refer to the README of the Scalar Envoy chart.

Configure custom values for Scalar Envoy chart

The Scalar Envoy chart is used via other charts (scalardb, scalardb-cluster, scalardl, and scalardl-audit), so you don’t need to create a custom values file for the Scalar Envoy chart. If you want to configure Scalar Envoy, you need to add the envoy.* configuration to the other charts.

For example, if you want to configure the Scalar Envoy for ScalarDB Server, you can configure some Scalar Envoy configurations in the custom values file of ScalarDB as follows.

  • Example (scalardb-custom-values.yaml) 
    envoy:
  configurationsForScalarEnvoy: 
    ...
  
scalardb:
  configurationsForScalarDB: 
     ...

Required configurations

Service configurations

You must set envoy.service.type to specify the Service resource type of Kubernetes.

If you accept client requests from inside of the Kubernetes cluster only (for example, if you deploy your client applications on the same Kubernetes cluster as Scalar products), you can set envoy.service.type to ClusterIP. This configuration doesn’t create any load balancers provided by cloud service providers.

envoy:
  service:
    type: ClusterIP

If you want to use a load balancer provided by a cloud service provider to accept client requests from outside of the Kubernetes cluster, you need to set envoy.service.type to LoadBalancer.

envoy:
  service:
    type: LoadBalancer

If you want to configure the load balancer via annotations, you can also set annotations to envoy.service.annotations.

envoy:
  service:
    type: LoadBalancer
    annotations:
      service.beta.kubernetes.io/aws-load-balancer-internal: "true"
      service.beta.kubernetes.io/aws-load-balancer-type: "nlb"

Optional configurations

If you want to control pod resources using the requests and limits of Kubernetes, you can use envoy.resources.

You can configure them using the same syntax as the requests and limits of Kubernetes. So, please refer to the official document Resource Management for Pods and Containers for more details on the requests and limits of Kubernetes.

envoy:
  resources:
    requests:
      cpu: 1000m
      memory: 2Gi
    limits:
      cpu: 2000m
      memory: 4Gi

If you want to control pod deployment using the affinity and anti-affinity of Kubernetes, you can use envoy.affinity.

You can configure them using the same syntax as the affinity of Kubernetes. So, please refer to the official document Assigning Pods to Nodes for more details on the affinity configuration of Kubernetes.

envoy:
  affinity:
    podAntiAffinity:
      preferredDuringSchedulingIgnoredDuringExecution:
        - podAffinityTerm:
            labelSelector:
              matchExpressions:
                - key: app.kubernetes.io/name
                  operator: In
                  values:
                    - scalardb-cluster
                - key: app.kubernetes.io/app
                  operator: In
                  values:
                    - envoy
            topologyKey: kubernetes.io/hostname
          weight: 50

If you want to monitor Scalar Envoy pods using kube-prometheus-stack, you can deploy a ConfigMap, a ServiceMonitor, and a PrometheusRule resource for kube-prometheus-stack using envoy.grafanaDashboard.enabled, envoy.serviceMonitor.enabled, and envoy.prometheusRule.enabled.

envoy:
  grafanaDashboard:
    enabled: true
    namespace: monitoring
  serviceMonitor:
    enabled: true
    namespace: monitoring
    interval: 15s
  prometheusRule:
    enabled: true
    namespace: monitoring

If you want to set SecurityContext and PodSecurityContext for Scalar Envoy pods, you can use envoy.securityContext and envoy.podSecurityContext.

You can configure them using the same syntax as SecurityContext and PodSecurityContext of Kubernetes. So, please refer to the official document Configure a Security Context for a Pod or Container for more details on the SecurityContext and PodSecurityContext configurations of Kubernetes.

envoy:
  podSecurityContext:
    seccompProfile:
      type: RuntimeDefault
  securityContext:
    capabilities:
      drop:
        - ALL
    runAsNonRoot: true
    allowPrivilegeEscalation: false

If you want to change the image repository and version, you can use envoy.image.repository to specify the container repository information of the Scalar Envoy container image that you want to pull.

envoy:
  image:
    repository: <SCALAR_ENVOY_CONTAINER_IMAGE>

If you’re using AWS or Azure, please refer to the following documents for more details:

TLS configurations (optional based on your environment)

You can enable TLS in:

  • Downstream connections between the client and Scalar Envoy.
  • Upstream connections between Scalar Envoy and Scalar products.

Enable TLS in downstream connections

You can enable TLS in downstream connections by using the following configurations:

envoy:
  tls:
    downstream:
      enabled: true
      certChainSecret: "envoy-tls-cert"
      privateKeySecret: "envoy-tls-key"

In this case, you have to create secret resources that include private key and certificate files for Scalar Envoy as follows:

kubectl create secret generic envoy-tls-cert --from-file=cert-chain=/path/to/your/certificate/file -n <NAMESPACE>
kubectl create secret generic envoy-tls-key --from-file=private-key=/path/to/your/private/key/file -n <NAMESPACE>

For more details on how to prepare private key and certificate files, see How to create private key and certificate files for Scalar products.

Enable TLS in upstream connections

You can enable TLS in upstream connections by using the following configurations:

envoy:
  tls:
    upstream:
      enabled: true
      overrideAuthority: "cluster.scalardb.example.com"
      caRootCertSecret: "scalardb-cluster-tls-ca"

In this case, you have to create secret resources that include CA certificate files as follows. You must set the root CA certificate file based on the upstream that you use (ScalarDB Cluster, ScalarDL Ledger, or ScalarDL Auditor).

  • ScalarDB Cluster

    kubectl create secret generic scalardb-cluster-tls-ca --from-file=ca-root-cert=/path/to/root/ca/cert/file/for/scalardb-cluster -n <NAMESPACE>

  • ScalarDL Ledger

    kubectl create secret generic scalardl-ledger-tls-ca --from-file=ca-root-cert=/path/to/root/ca/cert/file/for/scalardl-ledger -n <NAMESPACE>

  • ScalarDL Auditor

    kubectl create secret generic scalardl-auditor-tls-ca --from-file=ca-root-cert=/path/to/root/ca/cert/file/for/scalardl-auditor -n <NAMESPACE>

For more details on how to prepare private key and certificate files, see How to create key and certificate files for Scalar products.

Also, you can set the custom authority for TLS communication by using envoy.tls.upstream.overrideAuthority. This value doesn’t change what host is actually connected. This value is intended for testing but may safely be used outside of tests as an alternative to DNS overrides. For example, you can specify the hostname presented in the certificate chain file that you set by using scalardbCluster.tls.certChainSecret, ledger.tls.certChainSecret, or auditor.tls.certChainSecret, depending on which product you’re using. Envoy uses this value for verifying the certificate of the TLS connection with ScalarDB Cluster or ScalarDL.

Replica configurations (Optional based on your environment)

You can specify the number of replicas (pods) of Scalar Envoy using envoy.replicaCount.

envoy:
  replicaCount: 3

Taint and toleration configurations (Optional based on your environment)

If you want to control pod deployment by using the taints and tolerations in Kubernetes, you can use envoy.tolerations.

You can configure taints and tolerations by using the same syntax as the tolerations in Kubernetes. For details on configuring tolerations in Kubernetes, see the official Kubernetes documentation Taints and Tolerations.

envoy:
  tolerations:
    - effect: NoSchedule
      key: scalar-labs.com/dedicated-node
      operator: Equal
      value: scalardb