ODOCK.AI
Self Host

Stack Components

Stack Components

The self-hosted Odock stack is composed of a small number of core services and an optional monitoring layer.

This page describes the main parts of the reference stack at a conceptual level, not as an installation guide.

Core Platform Services

ServiceRole
postgresPersistent database for configuration, organisation data, and usage records.
redisRuntime support service used by the gateway.
traefikEntry point and reverse proxy for the API, UI, and optional Grafana routes.
serverOdock gateway runtime that handles requests and exposes the API.
ui-migrateOne-shot migration service that prepares the database schema.
uiOperator-facing web interface for managing the deployment.

Why The Stack Is Split This Way

The split between these services reflects common platform responsibilities:

  • request handling and governance at the gateway layer
  • operator workflows in the UI layer
  • persistent state in the data layer
  • ingress and exposure in the routing layer
  • operational support in the runtime layer

This makes it easier for teams to understand which parts are business-critical, which parts are infrastructure-facing, and which parts are optional for an early evaluation.

Optional Monitoring Services

A self-hosted deployment can also include an observability layer with components such as:

  • prometheus
  • alertmanager
  • loki
  • tempo
  • otel-collector
  • promtail
  • grafana
  • node-exporter
  • cadvisor
  • redis-exporter
  • postgres-exporter

These services provide platform-level telemetry for the gateway and the surrounding infrastructure.

Reference Deployment Model

The current public documentation describes a reference stack based on containerized services. That reference model is useful for evaluation because it makes the platform boundaries clear and easy to discuss with infrastructure teams.

If your team is considering self-hosting, contact us and we can explain how this reference architecture maps to your own environment.

On this page