Managed System Add-Ons - Rafay Product Documentation
Overview
Managed System Add-ons can be added to a cluster blueprint and be deployed on clusters. Once enabled in a blueprint, the required k8s software components and configuration are automatically deployed to the target clusters. Support and technical assistance is provided for the Managed Add-ons and Services.
| Managed System Addons | Components | Supported Distros |
|---|---|---|
| Managed Storage | Rook-Ceph | Upstream k8s |
| Managed Storage | GlusterFS (Deprecated) | Upstream k8s |
| Ingress Controller | Ingress Nginx | All |
| Monitoring & Alerting | Prometheus | All |
| Monitoring & Alerting | Metrics Server | All |
| Monitoring & Alerting | Alert Manager | All |
| Secrets Store CSI Driver | Secret Store CSI driver + AWS Secrets Manager provider specific plugins | Amazon EKS |
| Node Termination Handler | AWS Node Termination Handler (Deprecated) | Amazon EKS |
Managed Services
Managed Services are turnkey, high value integrations that can be based on either open source or purpose built software. Once enabled, the required k8s software components and configuration are automatically deployed to the target clusters. Support and technical assistance is provided for Managed Services.
| Managed Services | Components | Supported Distros |
|---|---|---|
| Monitoring & Alerting | Prometheus | All |
| Monitoring & Alerting | Metrics Server | All |
| Policy Management | OPA Gatekeeper | All |
| Backup & Restore | Velero | All |
| Network Policy | Cilium | All |
Mandatory Addons
The k8s Management Operator is selected by default and users are not allowed to disable this add-on. This is required for centralized management of Kubernetes clusters.
Optional Addons
Managed Storage
Managed Storage enables the Ceph/Rook storage addon, which can be deployed to your cluster(s) during provision to avail distributed storage systems.
Ingress Controller
Ingress controller, selected by default in the Controller, is a specialized load balancer for Kubernetes (and other containerized) environments. This specific add-on accepts traffic from outside the Kubernetes platform and load balance it to pods (containers) running inside the platform.
Monitoring & Alerting
Monitoring & Alerting, selected by default, collects the various metrics from different system services and generates alerts. This specific add-on helps to record real-time metrics in a time series database, with real-time alerting. Users are allowed to customize the Monitoring & Alerting add-on.
Customize Prometheus Add-On
Enabling Monitoring & Alerting Add-ons helps to update the controller dashboards of various metrics associated with clusters.
- Kube State Metrics: Exposes Prometheus format metrics based on the current state of the Kubernetes resources.
- Node Exporter: Exposes Prometheus format metrics based on the current state of the Kubernetes nodes.
- Helm Exporter: Exposes helm release, chart, and version statistics in the controller.
Kube State Metrics, Node Exporter, and Helm Exporter plays a major role in exposing the metrics in the controller dashboard. Invalid data might break the dashboard.
Prometheus Adapter: Prometheus Adapter is used for HPA (Horizontal Pod Autoscaler). Enabling this metric to use for automatically scaling workload up or down based on resource usage.
Metrics Server: Metrics Server provides metrics for resource utilization like CPU & Memory.
GPU Operator: To install the GPU Operator and have the dashboards work, users must specify the resource type as either Service or Pod, and add the required key-value labels.
- Resources Limits (Optional): Specify the number of resources a Container needs. Users can set the CPU and Memory Limits for the resources.
Click Save to apply all the changes made to the Monitoring & Alerting add-ons.
Secrets Store CSI Driver
During the pod start/restart, Secrets Store CSI Driver communicates with the provider using gRPC to retrieve the secret content from the external Secrets Store specified in the SecretProviderClass custom resource.
- Click the Wrench icon to customize the CSI Driver parameters.
Node Termination Handler (Amazon EKS only)
The AWS Node Termination Handler handles EC2 interruption events such as Spot interruptions, scheduled maintenance, instance terminations, and other lifecycle events.