The current LangSmith Helm chart (langchain-ai/helm) does not support configuring multiple hostnames through Helm values for the following ingress scenarios:
- Multiple ingress resources - Cannot create separate ingress configurations for different endpoints
- Differentiated access patterns - Cannot apply different security policies per hostname
These configurations are essential for production deployments where organizations need to separate human-facing and API-facing endpoints with different authentication.
Use Case
In our production environment, we need to expose LangSmith through two different hostnames for different access patterns:
- Human-facing endpoint (langsmith.company.com) - Web UI access with SSO authentication
- API-facing endpoint (langsmith-api.company.com) - Automated systems with API key authentication