Operating DIAL

This section is for DevOps and platform engineers who install, configure, and run DIAL. It covers local setups with Docker Compose, cloud deployment with Helm, model connections, configuration, authentication, observability, and production hardening. It assumes you know Docker for local setups and Kubernetes for cluster deployments.

Install DIAL

• Local setup — run a complete DIAL stack on one machine with Docker Compose, connected to an application, Ollama, vLLM, Azure OpenAI, or a remote DIAL Core
• Cloud deployment — deploy the Helm charts to AWS, Azure, GCP, or another Kubernetes cluster, then add Custom Apps, Quick Apps, or DIAL Admin

Connect models

Deploy models on a cloud provider so that DIAL can route requests to them.

• Amazon Bedrock
• Azure OpenAI
• Vertex AI
• Databricks

Configure and secure

• Configuration — configuration files, environment variables, and Docker Compose settings
• Auth and access control — API keys, JWT, roles, rate limits, and SSO providers

Run in production

• Observability — tracing, metrics, logging, alerting, and analytics setup
• Production readiness — availability, scaling, secrets, backup, upgrades, cost, and hardening
• Troubleshooting — map a symptom or an HTTP status code from DIAL Core to its cause and resolution