Releases

Issue 1: Deployment process is stuck or failing to complete.

Fix:

Ensure your internet connection is stable.
Double-check credentials (API and SSH keys) in GPU Container settings.
Navigate to specific GPU container’s logs for error messages. Send logs to support team to provide solution for identified issues.
If issues still persist after following the steps above, delete the failed deployment and try again.

Issue 2: No GPU instance available for your workflow .

Fix:

Verify the availability of the requested GPU instance on the Kaisar platform.
Increase or decrease the resource requirements (e.g., GPU cores, memory) to find the most closest GPU instance for deployment.
Verify whether you have enough tokens in your connected crypto wallet or KAI tokens in your Kaisar account.
Try switching to a different pricing model and GPU instance and try deployment again.
If troubleshooting steps above do not resolve your issue, contact support via the methods outlined in the Support section of this documentation. .

Issue 3: Not able to connect to GPU container.

Fix:

Ensure exposed ports for GPU container allow outbound and inbound traffic .
Verify credentials (API and SSH public keys) are correctly configured .
Try SSHing the container to ensure it is reachable.
If your organization makes use of a firewall, try disabling it or use a VPN to establish connection with GPU container.

Issue 4: Slow performance or high latency when leveraging GPU workflows.

Fix:

Monitor resource (CPU, GPU, memory, and network) usage to identify performance and latency issues. To view resource usage navigate to the container details page and click on the Logs tab.
Consider upgrading to a higher-spec GPU container from a Kaisar list of available instances.
Limit and optimize GPU usage by running tasks more effectively and reducing concurrent execution.

Last updated 7 months ago

Issue 1: Deployment process is stuck or failing to complete.

Fix:

Ensure your internet connection is stable.
Double-check credentials (API and SSH keys) in GPU Container settings.
Navigate to specific GPU container’s logs for error messages. Send logs to support team to provide solution for identified issues.
If issues still persist after following the steps above, delete the failed deployment and try again.

Issue 2: No GPU instance available for your workflow .

Fix:

Verify the availability of the requested GPU instance on the Kaisar platform.
Increase or decrease the resource requirements (e.g., GPU cores, memory) to find the most closest GPU instance for deployment.
Verify whether you have enough tokens in your connected crypto wallet or KAI tokens in your Kaisar account.
Try switching to a different pricing model and GPU instance and try deployment again.
If troubleshooting steps above do not resolve your issue, contact support via the methods outlined in the Support section of this documentation. .

Issue 3: Not able to connect to GPU container.

Fix:

Ensure exposed ports for GPU container allow outbound and inbound traffic .
Verify credentials (API and SSH public keys) are correctly configured .
Try SSHing the container to ensure it is reachable.
If your organization makes use of a firewall, try disabling it or use a VPN to establish connection with GPU container.

Issue 4: Slow performance or high latency when leveraging GPU workflows.

Fix:

Monitor resource (CPU, GPU, memory, and network) usage to identify performance and latency issues. To view resource usage navigate to the container details page and click on the Logs tab.
Consider upgrading to a higher-spec GPU container from a Kaisar list of available instances.
Limit and optimize GPU usage by running tasks more effectively and reducing concurrent execution.

Last updated 7 months ago