Containers and Kubernetes

Users that do not have cluster administrator privileges can install and run the operator by passing a custom service account with operator privileges to the serviceAccountNameOverride Helm chart parameter.

Operator helm chart privileges with service accounts

Users that do not have cluster administrator privileges can install and run the operator by passing a custom service account with operator privileges to the serviceAccountNameOverride Helm chart parameter. The serviceAccountNameOverride parameter prevents the Helm chart from creating a default service account for the operator installation.

The cluster administrator must create the custom service account and grant it operator privileges before the non-admin user installs the Helm chart.

For details about granting operator privileges, see Installing the Vertica DB operator. For details about serviceAccountNameOverride, see Helm chart parameters.

Install the helm chart without the admission controller webhook

Users with limited permissions might want to install only the operator to manage Vertica on Kubernetes. The webhook.enable Helm chart parameter controls whether the Helm chart installs the admission controller webhook along with the operator.

For detailed installation steps, see Installing the Vertica DB operator. For details about webhook.enable, see Helm chart parameters.

Managed Kubernetes services support

Vertica supports managed Kubernetes services on Azure Kubernetes Service (AKS).

Air-gapped operator with helm chart

As a security feature, you can deploy the operator in an air-gapped cluster, which is a cluster that is disconnected from other networks. Use the following parameters to configure the isolated operator:

  • image.repo

  • rbac_proxy_image.name

  • rbac_proxy_image.repo

  • imagePullSecrets

For additional details about each of these parameters, see Helm chart parameters.

Customize LoadBalancer service types

Assign a static IP address for subclusters that use the LoadBalancer service type with the subcluster[i].loadBalancerIP parameter. This is useful for bare metal Vertica on Kubernetes deployments.

Add custom annotations to implementation-specific services with the subclusters[i].serviceAnnotations parameter. Managed Kubernetes offerings use annotations on service objects to control logging and other actions.

For details about each parameter, see Custom resource definition parameters. For a list of supported Managed services, see Containerized environments.

Configure online upgrade traffic redirect wait time

During an upgrade, the operator runs a reconcile iteration and requeues any resource objects for additional work if their actual state does not match the desired state. Set the upgradeRequeueTime parameter to determine how long the operator waits to complete any requeued work.

For additional details, see Upgrading Vertica on Kubernetes and Custom resource definition parameters.

VerticaAutoscaler custom resource

The new VerticaAutoscaler custom resource automatically scales existing subclusters by subcluster or pod. You can select the metric and trigger that determines when the workload resource scales.

For details, see VerticaAutoscaler custom resource.

Prometheus integration

Vertica on Kubernetes integrates with Prometheus to collect metrics on the VerticaDB operator. Helm chart parameters configure role-based access control (RBAC) or allow external client connections without RBAC authorization.

For details, see Prometheus integration.