Guides

Managing GPU Instances

Learn how to create, monitor, and scale your GPU instances to suit your project needs on GPUiq.

const metrics = await fetch('https://api.example.com/v1/instances/{instance_id}/metrics', {
  headers: { 'Authorization': `Bearer ${YOUR_API_KEY}` }
}).then(r => r.json());

{
  "gpu_util": 75.2,
  "vram_used_gb": 23.5,
  "temperature_c": 72,
  "timestamp": "2024-10-15T12:00:00Z"
}

[
  {"id": "snap-123", "name": "pre-training-backup", "created": "2024-10-15T10:00:00Z"}
]

Creating and Customizing Instances

Create GPU instances through the GPUiq dashboard or API to match your compute needs, such as AI training or 3D rendering. Select GPU types like NVIDIA A100 or H100, configure CPU/RAM, and set storage volumes.

Navigate to https://dashboard.example.com/instances.

Choose Configuration

Select GPU model, vCPU count (e.g., 8-64), RAM (16-512 GB), and storage (100 GB NVMe SSD minimum).

Deploy Instance

Click "Launch" and wait 2-5 minutes for provisioning.

Customize via API for automation:

const response = await fetch('https://api.example.com/v1/instances', {
  method: 'POST',
  headers: { 'Authorization': `Bearer ${YOUR_API_KEY}`, 'Content-Type': 'application/json' },
  body: JSON.stringify({
    name: 'ai-training-node',
    gpu_type: 'A100',
    vcpu: 16,
    memory_gb: 128,
    storage_gb: 500
  })
});
const instance = await response.json();
console.log(instance.id);

curl -X POST https://api.example.com/v1/instances \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "ai-training-node",
    "gpu_type": "A100",
    "vcpu": 16,
    "memory_gb": 128,
    "storage_gb": 500
  }'

import requests
response = requests.post(
  'https://api.example.com/v1/instances',
  headers={'Authorization': 'Bearer YOUR_API_KEY', 'Content-Type': 'application/json'},
  json={
    'name': 'ai-training-node',
    'gpu_type': 'A100',
    'vcpu': 16,
    'memory_gb': 128,
    'storage_gb': 500
  }
)
instance = response.json()
print(instance['id'])

Start with smaller configs to test, then scale up. GPUiq auto-scales storage on demand.

Monitoring Resource Usage and Performance

Track GPU utilization, memory, and network I/O in real-time to optimize costs and performance.

Use the dashboard for visuals or API for integrations.

Go to https://dashboard.example.com/instances/`{instance_id}`/metrics.

View charts for:

GPU utilization (<80% ideal for cost savings)
VRAM usage
Temperature (<85°C)

Set alerts for high usage.

Instance Lifecycle Management

Control your instances efficiently to minimize costs.

Stop

Pause billing while preserving data. Resume instantly.

Resume

Restart from stopped state in <2 minutes.

Delete

Permanently remove. Use snapshots first for backups.

Stop and Resume

path

instance_idstring

Required

Instance identifier from creation response.

Stop via API:

curl -X POST https://api.example.com/v1/instances/{INSTANCE_ID}/stop \
  -H "Authorization: Bearer YOUR_API_KEY"

Resume similarly with /resume.

Delete

Confirm deletion to avoid data loss.

Using Snapshots for Backups and Restores

Snapshots capture instance state for backups, migrations, or quick restores. Create before stopping or scaling.

Snapshots incur minimal storage costs and enable up to 70% savings by reusing configs.

Was this page helpful?

Built with Documentation.AI

Last updated today