Cluster Management

The Clusters section of the dashboard provides comprehensive tools for creating, configuring, and managing computing clusters in the Manta platform.

Note

Visual guides and screenshots will be added in future documentation updates.

Overview

Clusters are the fundamental organizational unit in Manta, grouping computing nodes that work together on distributed tasks. Through the dashboard, you can:

  • Create new clusters with custom configurations

  • Monitor cluster health and resource utilization

  • Manage node membership and permissions

  • Configure network and security settings

  • Track cluster activity and performance metrics

Creating a Cluster

Step-by-Step Process

  1. Navigate to the Clusters section

  2. Click “Create New Cluster” button

  3. Fill in the cluster configuration form

  4. Review and confirm settings

  5. Save generated access tokens

Configuration Options

  • Basic Information - Cluster name (unique identifier) - Description and purpose - Tags for organization

  • Resource Settings - Maximum node count - Resource quotas (CPU, memory, storage) - Scheduling policies

  • Network Configuration - Communication protocols - Port assignments - Security settings

  • Access Control - User permissions - API access tokens - Role assignments

Cluster Dashboard

Each cluster has its own detailed dashboard showing:

Overview Tab

  • Cluster status and health

  • Node count and distribution

  • Active swarms and tasks

  • Resource utilization graphs

Nodes Tab

  • List of connected nodes

  • Node status and resources

  • Connection history

  • Performance metrics

Activity Tab

  • Recent events and actions

  • Audit log

  • Error reports

  • System notifications

Settings Tab

  • Configuration management

  • Token regeneration

  • Permission updates

  • Advanced options

Managing Clusters

Common Operations

  • Start/Stop: Control cluster availability

  • Scale: Add or remove node capacity

  • Update: Modify configurations

  • Archive: Preserve completed experiments

  • Delete: Remove unused clusters

Monitoring Tools

  • Real-time resource graphs

  • Node distribution maps

  • Task allocation views

  • Performance trends

Maintenance Actions

  • Health checks

  • Resource rebalancing

  • Log collection

  • Backup operations

Best Practices

Cluster Organization

  • Use descriptive names

  • Apply consistent tagging

  • Document cluster purposes

  • Regular cleanup of unused clusters

Resource Management

  • Set appropriate quotas

  • Monitor utilization trends

  • Plan for peak loads

  • Implement auto-scaling rules

Security Considerations

  • Rotate tokens regularly

  • Limit access permissions

  • Enable audit logging

  • Monitor suspicious activity

Next Steps