Skip to main content

Automated Deployment (Recommended)

The deployment script automates:

  • Kubernetes secret creation
  • Helm repository configuration
  • LiteLLM Proxy deployment
  • Integration secret creation for CodeMie

Pre-Deployment Checklist​

Before running the script, ensure you have completed the following configuration steps:

  1. Clone the Repository: Clone the codemie-helm-charts repository and navigate to its root directory.

  2. Configure API and Proxy Values: Update both the codemie-api/values-<cloud>.yaml and litellm/values-<cloud_name>.yaml files with your environment-specific settings as described in Configure Values.

  3. Configure LLM Models: Review and customize your desired models and regions in the litellm/config/litellm-<cloud>-config.yaml file, as detailed in Model Configuration.

warning

Completing all configuration steps mentioned in the checklist is mandatory for a successful installation.

Run the Deployment Script​

Once all configurations are in place, execute the following command from the root of the repository. Replace <cloud_name> with your target cloud (aws, azure, or gcp) and specify the desired version.

bash helm-charts.sh --cloud <cloud_name> --version=x.y.z --mode all --optional litellm

During execution, the script may prompt you to enter authentication details if you are deploying for Azure or GCP.

Next Steps​

After successful deployment:

  • Verify the deployment by checking pod status: kubectl get pods -n litellm
  • Access the LiteLLM UI using the configured domain
  • For migration from DIAL Proxy, see Migration Guide