Azure
Run Agent Gateway on Azure to leverage Azure OpenAI, AKS, and Azure services.
Deployment options
Azure Kubernetes Service (AKS)
For AKS deployments, use kgateway which provides native Kubernetes Gateway API support, dynamic configuration, and MCP service discovery.
Azure Container Apps
Run Agent Gateway as a serverless container.
az containerapp create \
--name agentgateway \
--resource-group my-rg \
--environment my-env \
--image ghcr.io/agentgateway/agentgateway:latest \
--target-port 3000 \
--ingress external \
--secrets openai-key=secretref:openai-api-key \
--env-vars AZURE_OPENAI_ENDPOINT=https://my-resource.openai.azure.comAzure integrations
| Integration | Purpose |
|---|---|
| Azure OpenAI | Access GPT-4 and other models |
| Azure Key Vault | Secure API key storage |
| Azure Application Gateway | Load balancing with WAF |
| Azure Monitor | Logs and metrics |
| Azure Application Insights | Distributed tracing |
Azure role assignments
Assign roles to the managed identity:
# Get the managed identity principal ID
PRINCIPAL_ID=$(az identity show --name agentgateway-identity \
--resource-group my-rg --query principalId -o tsv)
# Grant Azure OpenAI access
az role assignment create \
--assignee $PRINCIPAL_ID \
--role "Cognitive Services OpenAI User" \
--scope /subscriptions/<sub-id>/resourceGroups/my-rg/providers/Microsoft.CognitiveServices/accounts/my-openai
# Grant Key Vault access
az role assignment create \
--assignee $PRINCIPAL_ID \
--role "Key Vault Secrets User" \
--scope /subscriptions/<sub-id>/resourceGroups/my-rg/providers/Microsoft.KeyVault/vaults/my-vault