Skip to content

Job Management UI

The Jobs section of the OICM+ platform allows you to create, monitor, and manage jobs that consume computing resources for tasks like machine learning model training.

1. Initiating a New Job

  1. Navigate to the Jobs section.
  2. Create a new job by clicking + New Job.
  3. Fill in the form:
  4. Title – Descriptive name of your job.
  5. Job Type – Choose the framework (e.g., PyTorch, Ray, TensorFlow).
  6. Tags – Add optional labels for organization and filtering.

Job Creation

Once created, your job appears in the Jobs list alongside existing entries.

Jobs List

2. Job-Specific Page

Click a job in the list to view its dedicated interface, featuring multiple tabs for detailed management:

Job Specific Page

2.1 Scripts

  • Upload Options – Single file or directory upload.
  • Edit Files – Make changes directly in the UI.
  • Delete & Refresh – Remove files or refresh the file list.

Note: For multiple files, name your main script main.py and config file config.yaml to help the system identify primary execution files.

2.2 Workers

  • Compute Units – View all allocated workers.
  • Status & Resources – Track each worker’s resource usage and allocation.
  • Detailed Specs – Inspect hardware configuration and specifications.

2.3 Logs

  • Real-Time Output – Monitor console logs for each worker.
  • Worker Dropdown – Select a worker to see its specific logs.
  • Live Updates – Follow the execution flow in real time.

2.4 Events

  • Chronological Timeline – Capture all job-related events in order.
  • Notifications – Highlight errors, warnings, or major transitions.
  • Diagnostics – Help you pinpoint issues and track job states.

2.5 System Metrics

  • GPU Indicators – Core utilization, temperature, memory usage.
  • Real-Time & Historical – Interactive graphs for current or past performance.
  • Hardware Usage – Encoder/decoder utilization and more.

2.6 Settings

  • Update Configuration – Modify job title, type, and tags.
  • Delete Job – Remove the entire job from the platform when no longer needed.

3. Monitoring Job Performance

Use tabs in combination for a full picture:

  • Events + System Metrics – Correlate timeline events with resource usage to detect performance drops.
  • Workers + System Metrics – Identify bottlenecks by comparing worker status with CPU/GPU usage.
  • Logs + Events – Debug issues by viewing runtime output alongside relevant warnings or errors.

4. Best Practices

  • Assign Resources Properly – Check the Workers tab to ensure correct CPU/GPU allocations.
  • Track System Metrics – Optimize usage by monitoring GPU memory, temperature, and utilization.
  • Review Events – Understand the lifecycle and troubleshoot issues quickly.
  • Use Logs – Drill down into specific worker logs for real-time debugging.
  • Tag Jobs – Organize and filter for better discoverability, especially in large teams.

Next Steps

  • Jobs Overview – Explore the underlying concepts of job management.
  • Jobs Examples – Learn about writing scripts for advanced job configurations.
  • Resource Allocation – Manage CPU/GPU resources across multiple jobs effectively.