Installation
Palette is available as a self-hosted application that you install in your environment. The self-hosted version is a dedicated Palette environment hosted on VMware instances or in an existing Kubernetes cluster. Palette is available in the following modes:
| Supported Platform | Description | Install Guide |
|---|---|---|
| VMware | Install Palette in VMware environment. | Install on VMware |
| Kubernetes | Install Palette using a Helm Chart in an existing Kubernetes cluster. | Install on Kubernetes |
Airgap Installation
You can also install Palette in an airgap environment. For more information, refer to the Airgap Installation section.
| Supported Airgap Platform | Description | Install Guide |
|---|---|---|
| VMware | Install Palette in VMware environment using your own OCI registry server. | VMware Airgap Install |
| Kubernetes | Install Palette using a Helm Chart in an existing Kubernetes cluster with your own OCI registry server OR use AWS ECR. | Kubernetes Airgap Install |
The next sections provide sizing guidelines we recommend you review before installing Palette in your environment.
Size Guidelines
This section lists resource requirements for Palette for various capacity levels. In Palette VerteX, the terms small, medium, and large are used to describe the instance size of worker pools that Palette is installed on. The following table lists the resource requirements for each size.
The recommended maximum number of deployed nodes and clusters in the environment should not be exceeded. We have tested the performance of Palette with the recommended maximum number of deployed nodes and clusters. Exceeding these limits can negatively impact performance and result in instability. The active workload limit refers to the maximum number of active nodes and pods at any given time.
| Size | Nodes | CPU | Memory | Storage | MongoDB Storage Limit | MongoDB Memory Limit | MongoDB CPU Limit | Total Deployed Nodes | Deployed Clusters with 10 Nodes |
|---|---|---|---|---|---|---|---|---|---|
| Small | 3 | 8 | 16 GB | 60 GB | 20 GB | 4 GB | 2 | 1000 | 100 |
| Medium (Recommended) | 3 | 16 | 32 GB | 100 GB | 60 GB | 8 GB | 4 | 3000 | 300 |
| Large | 3 | 32 | 64 GB | 120 GB | 80 GB | 12 GB | 6 | 5000 | 500 |
Instance Sizing
| Configuration | Active Workload Limit |
|---|---|
| Small | Up to 1000 Nodes each with 30 Pods (30,000 Pods) |
| Medium (Recommended) | Up to 3000 Nodes each with 30 Pods (90,000 Pods) |
| Large | Up to 5000 Nodes each with 30 Pods (150,000 Pods) |
Proxy Requirements
Palette connects to the internet to download images and packages. If your environment uses a proxy server, ensure the following domains and ports are accessible. The proxy server should meet the following requirements:
-
A proxy used for outgoing connections should support both HTTP and HTTPS traffic.
-
Allow connectivity to domains and ports in the table.
-
Review the gRPC and Proxies page to learn more about Palette's support for gRPC in a proxy environment.
Top-Level Domain Port Description spectrocloud.com 443 Spectro Cloud content repository and pack registry s3.amazonaws.com 443 Spectro Cloud VMware OVA files gcr.io 443 Spectro Cloud and common third party container images ghcr.io 443 Kubernetes VIP images docker.io 443 Common third party content googleapis.com 443 For pulling Spectro Cloud images docker.com 443 Common third party container images raw.githubusercontent.com 443 Common third party content projectcalico.org 443 Calico container images quay.io 443 Common third party container images grafana.com 443 Grafana container images and manifests github.com 443 Common third party content k8s.gcr.io 443 Kubernetes images [deprecated] registry.k8s.io 443 Kubernetes images docker.pkg.dev 443 Common third party content