Skip to main content

Installation Overview

Once your cloud infrastructure or on-premise server configuration is complete according to the prerequisites, you can install all EVA components using Helm.

EVA consists of EVA App, EVA Agent, and EVA Vision based on their respective roles, and each component is provided as an independent Helm Chart.

To ensure a smooth operation, please follow the installation process in the order described below.

🔔 An AWS Access Key is strictly required for authentication and access to the Mellerikat Cloud Image Registry (ECR) to proceed with the installation.



Installation Process​

Since the EVA system has dependencies between its components, we recommend the following step-by-step installation:

  1. Infrastructure Dependencies: First, you must establish the underlying infrastructure, including the database (Qdrant) and the inference server (vLLM) required for the EVA Agent.
  2. EVA Agent Installation: Once the dependencies are configured, install the Agent responsible for AI inference and data processing.
  3. EVA Vision Installation: Install the module responsible for vision analysis of video streams.
  4. EVA App Installation: Finally, install the web interface and management dashboard that users will access.


Maintenance and Uninstallation​

You can easily manage the lifecycle of EVA through Helm.

  • Update: After modifying settings in values.yaml or upgrading versions, apply the changes using the helm upgrade command.
  • Uninstallation: You can remove deployed resources using the helm uninstall command. (Note: To prevent data loss, automatically created PVCs/PVs may need to be deleted manually.)


Detailed Guides​