Rental Management
Once you've successfully rented a GPU on Nebula AI, managing your session effectively ensures optimal performance, cost-efficiency, and uninterrupted workflow. This section covers everything you need to know about monitoring usage, extending rentals, handling early termination, and troubleshooting issues.
Monitoring Your Active Rentals
After confirming a rental, you can access your Rental Dashboard, which provides real-time tracking and management tools. From this interface, you can:
View GPU Status – Check if the rented GPU is online, in use, or idle.
Monitor Performance Metrics – Track VRAM usage, GPU load, power consumption, and network speed.
Check Remaining Rental Time – See how much time is left before the rental expires.
Access Connection Details – Get SSH login, Jupyter Notebook links, or API credentials.
Keeping an eye on resource usage ensures you're not paying for idle compute time and helps you adjust workloads accordingly.
Extending Your Rental Duration
If your workload requires more computing time, you can extend the rental before it expires.
In the Rental Dashboard, locate your active session and select "Extend Rental".
Choose the additional duration (hourly or long-term options).
Confirm and approve the transaction in your wallet.
Ending a Rental Session Early
Rentals automatically terminate at the end of the rental period, but you may end your session manually if you no longer need the GPU.
Navigate to the Rental Dashboard and select "End Rental".
Confirm the termination request.
The system will immediately release the GPU back to the marketplace.
Important:
No refunds are given for unused time if you terminate early.
All session data is erased upon termination, so back up your work before ending the rental.
Auto-Termination & Data Management
When a rental reaches its expiration time, the GPU is automatically released and reset. This means:
All session data is deleted – Ensure you have backups.
The GPU is immediately available for other renters.
Pending payments and transactions are settled automatically via smart contracts.
If you require persistent storage across rental sessions, Nebula AI will introduce cloud-linked storage solutions in a future update.
Handling GPU Downtime & Failures
While rare, there may be instances where your rented GPU goes offline or underperforms. This could be due to:
Host hardware failure
Network connectivity issues
Unexpected driver or software crashes
If a GPU failure occurs during an active rental, the system will:
Automatically detect downtime and attempt to restore functionality.
If the downtime persists, the session may be credited or refunded based on the issue.
Support tickets can be opened if manual intervention is needed.
Troubleshooting Common Issues
Issue
Cause
Solution
GPU not responding after rental starts
Temporary host delay
Wait a few minutes or restart the session
Slow performance
High concurrent load or network issue
Check host GPU specs and connection speed
Session disconnected unexpectedly
Network instability or host reboot
Reconnect using new session credentials
Cannot extend the rental
Insufficient funds or rental period limit reached
Ensure the wallet has enough $NAI or select a shorter extension
GPU unexpectedly terminated
Host failure or forced stop
Contact support for possible credit/refund
If persistent issues arise, reach out via Nebula AI support or community channels.
Last updated