Hi there, We have been trying to deploy Vapi via the AWS Marketplace but are coming across the same deployment issues with your Pulumi stacks. NodeGroups keep failing to create for the same reason:

create: unexpected state 'CREATE_FAILED', wanted target 'ACTIVE'. last error: eks-persistence-e4cb48a1-58c8-a28d-f7e3-942b2ec6001b: AsgInstanceLaunchFailures: Could not launch On-Demand Instances. Unsupported - The requested configuration is currently not supported. Please check the documentation for supported configurations. Launching EC2 instance failed.

Is there someone we can talk to in your technical team to resolve this? As your marketplace offering runs the Pulumi deployment in an EC2 has meant that we currently have a load of hanging infrastructure in a broken state. Ideally we would like to discuss other onprem hosting methods with your team. Thanks, Karik

Hi Karik, I'm sorry to hear about your AWS Marketplace deployment issues. Based on the error message, it appears there's an unsupported configuration causing the NodeGroups to fail during creation. For on-premises hosting alternatives, Vapi offers several options: - Dockerized deployment on any cloud provider using your own GPUs ([docs.vapi.ai/enterprise/onprem](https://docs.vapi.ai/enterprise/onprem)) - Client-hosted Vapi within your own clusters - Private AWS server deployment with your own infrastructure To help troubleshoot your current issue: Which AWS region and instance type are you attempting to deploy to? This information would help our technical team better understand the configuration issue. For direct assistance, please contact [email protected] to discuss both your current deployment issues and on-premises options. Sources: - [Vapi On-Prem Documentation](https://docs.vapi.ai/enterprise/onprem) - [Discord Channel Alternative hosting methods](https://discord.com/channels/1211482211119796234/1334521938478891104) - [Discord Channel Client-hosted options](https://discord.com/channels/1211482211119796234/1333396591255748651)

@Vapi Support Bot I would love to know more about teh Dockerized deployment on any cloud provider using your own GPUs. How do we set that up?

According to [Vapi On-Prem Documentation](https://docs.vapi.ai/enterprise/onprem), Vapi can be deployed in a dockerized format on any cloud provider using your own GPUs. The system automatically adjusts GPU resources based on your workload. While I can see that Vapi On-Prem keeps your audio and text data in your cloud and operates in VPCs you control, I don't have the specific setup instructions in my documentation. Which aspect of the Dockerized deployment are you most interested in - initial setup process, GPU requirements, or integration with your existing infrastructure? Source: - [Vapi On-Prem Documentation](https://docs.vapi.ai/enterprise/onprem)

Hey, really sorry for your experience. If possible, could you record a loom video describing this error/behavior or inconsistency in the dashboard with the steps to reproduce it? That way, I can share this with my team, and we'll take a look. We'll even try to reproduce the same so we can fix it for you. Also, I'll request you to share your organization ID and loom video in DM.

@Shubham Bajaj your bot suggested above that there are other on prem deployment options. Could you lay out what they are?

You can now deploy Vapi on your AWS server through the AWS Marketplace. Use this link to get started: