Add Cluster

Overview

The Add Cluster page allows administrators to manually define and configure a new Cluster in the DataFloem platform.

Using this interface, users can specify the cluster's general identity, execution mode, its dedicated Kubernetes Namespace, and detailed hardware limits for the cluster components (Engine, Master, and Slave).

Once created, the respective pods are provisioned on Kubernetes according to the selected operating mode (Dynamic/Static) and defined resource limits.

Prerequisites

Before creating a Cluster, ensure that the following conditions are met:

You must have Administrator (Admin) privileges.
The cluster name must be unique.
A valid Kubernetes Namespace must be available for assignment.
Ensure that the images to be used (Engine, Master, Slave) are accessible on the Docker Registry.

Access Requirements

To access this page, the user must either have Administrator privileges or the CREATE permission for the Clusters module.

The CREATE permission allows users to create new groups. Users without administrator privileges or the required permission cannot access this page.

If you do not have the required permission, ask your administrator to grant it from one of the authorization pages below.

Authorization

Users without the required authorization will receive the following error:

Access Denied

You don't have access to this resource.

Cluster Information

The following details must be configured under each tab when creating a cluster.

General Configuration

Defines the general identity of the cluster on Kubernetes and its behavioral rules.

Field	Required	Description
Name	Yes	Defines a unique name for the cluster.
Kubernetes Namespace	Yes	Determines the namespace where the cluster will run on Kubernetes.
Available To	No	Restricts which projects or user groups can access this cluster.
Configuration	No	Selects the default configuration template applicable to the cluster.
Idle Timeout	No	Specifies, in seconds, how long an unused cluster remains running before it is automatically shut down. The default value is `3600`; if left empty or set to `0`, the backend uses `3600` seconds.
Is Dinamic	Yes	Determines if the cluster is dynamic (auto-scaling) or static.
Globally Available	Yes	If enabled, the cluster becomes accessible to all systems and users.
Disable App Validation	Yes	Disables application validation steps for this cluster.
Is Active	Yes	Determines whether the cluster becomes active (usable) immediately after creation.

Engine Configuration

Configures the resource limits and environment parameters for the cluster's engine pod.

Field	Required	Description
Engine Disabled	Yes	If enabled, the engine service within the cluster is disabled.
Image	Yes	Selects the Docker image to be used for the Engine component.
Image Pull Secret Name	No	The Kubernetes Secret name required to pull images from a private registry.
Minimum / Maximum CPU	Yes	The lowest and highest CPU limits for the Engine pod.
Minimum / Maximum Memory	Yes	The lowest and highest RAM limits for the Engine pod.
Data Disk	Yes	Determines whether a Persistent Volume is attached to the pod.

Master Configuration

Defines resource limits and image parameters for the Master pod coordinating the cluster.

Field	Required	Description
Master Disabled	Yes	If enabled, the Master component of the cluster is disabled.
Image	Yes	Specifies the Docker image for the Master component.
Image Pull Secret Name	No	The Kubernetes Secret name required to pull the Master image.
Minimum / Maximum CPU	Yes	The minimum and maximum CPU allocated for the Master pod.
Minimum / Maximum Memory	Yes	The minimum and maximum RAM allocated for the Master pod.
Data Disk	Yes	Determines whether a Persistent Volume is assigned to the Master pod.

Slave (Worker) Configuration

Determines the number, images, and resource limits of the Slave pods executing the workload.

Field	Required	Description
Worker Count	Yes	Determines how many Slave (worker) pods will run in the cluster (Default: `1`).
Image	Yes	Defines the Docker image for the Slave pods.
Image Pull Secret Name	No	The Kubernetes Secret name used to pull the Slave image.
Minimum / Maximum CPU	Yes	The lower and upper CPU limits each Slave pod can consume.
Minimum / Maximum Memory	Yes	The lower and upper RAM limits each Slave pod can consume.
Data Disk	Yes	Determines whether persistent disk storage is enabled for Slave pods.

Cluster Creation Process

Follow these steps to create a cluster:

Open the Add Cluster modal.
In the General tab, specify the cluster name, namespace, and execution mode.
Configure resource and image settings in the Engine, Master, and Slave tabs based on requirements.
Set up Node Selectors and Tolerations if applicable.
Click the Save button to complete the process.

Once the cluster is successfully created:

The cluster is registered in the system with the specified configurations.
Pod provisioning begins under the designated Kubernetes namespace.
It will appear in the active list if Is Active was selected.

Default Behavior

When creating a cluster, the system applies the following default behaviors:

The Is Active field is active (true) by default.
Idle Timeout defaults to 3600 seconds.
Worker Count defaults to 1.
Engine Disabled, Master Disabled, and Data Disk switches are inactive (false) by default.

Validations

The following validations are performed before the cluster is created:

The cluster Name cannot be left blank and must be unique.
The Kubernetes Namespace field must be selected.
Images and resource limits (CPU/Memory) in the Engine, Master, and Slave tabs must be valid and cannot be empty.

Notes

Dynamic clusters can auto-scale (when Is Dinamic is active) based on workload or resource demands.
Ensure CPU (m) and Memory (Mi/Gi) units match Kubernetes standards to prevent scheduling errors.