DIAL Stack
This reference summarizes the technologies and infrastructure behind DIAL. For architects, Deployment profiles explains which parts to adopt; Architecture highlights explains their responsibilities.
Service technologies
| Service | Technology |
|---|---|
| DIAL Core | Java, Vert.x |
| DIAL Chat | TypeScript, Next.js |
| DIAL Admin Frontend | TypeScript, Next.js; connects directly to Core and optional service APIs |
| DIAL Analytics and Enrichment Runner | Java, Spring Boot |
| Model adapters, DIAL Log Parser, and selected application runtimes | Python; frameworks and dependencies vary by service |
| DIAL Scheduler | Python, DIAL SDK, APScheduler, SQLAlchemy |
The table identifies implementation families, not a shared runtime requirement for every service. Deployable component images carry their own dependencies; an application extension can introduce additional libraries or a different runtime.
Use the selected component release for exact runtime and dependency versions.
Backing infrastructure
| Dependency | Purpose |
|---|---|
| Blob storage | Durable resources; supported backends include AWS S3, Azure Blob Storage, Google Cloud Storage, compatible object stores, and filesystem storage |
| Valkey / Redis-compatible store | Cached resources, pending writes, and shared coordination state |
| Identity provider | Enterprise sign-in and identity claims; supported integrations use OIDC/OAuth 2.0 |
| Secret manager | Deployment credentials and keys |
| Observability stores | Selected metrics, trace, log, and analytics destinations |
| Optional component databases | Component-specific state, including Admin configuration, ClickHouse and PostgreSQL for analytics, and PostgreSQL for DIAL Scheduler |
Blob storage and Valkey serve Core's resource and coordination model. Identity, secrets, and observability infrastructure depend on the deployment's chosen integrations; optional component stores are additional dependencies, not replacements for Core's stores.
Core does not require a relational database. Optional services can require additional stores. See State and data lifecycle before choosing persistence and recovery settings.
Deployment
Docker Compose supports local evaluation. Kubernetes deployments use Helm charts and your selected cloud or self-managed infrastructure. See the Helm overview, identity-provider guides, and observability for configuration.
Further reading
- Services — find each component's dependencies and configuration
- Deployment profiles — which services and infrastructure each deployment profile needs
- Deploy DIAL to a cluster — choose a Helm installation guide
- Configure observability — connect services to your telemetry backends