Skip to content

Explore your domain in the Observatory

The Protean Observatory ships two interactive views of a running domain, the Domain Visualizer and the Timeline. The Domain Visualizer draws your aggregates and their relationships from the live IR, and the Timeline lets you browse, filter, and trace every message the system has processed.

Here is how to use both. For startup options and the full endpoint list see the protean observatory reference; for the tracing data behind the views see Correlation and causation IDs.

Start the Observatory and open it in a browser:

protean observatory --domain=myapp
# Visit http://localhost:8080

Visualize the domain topology

Open the Domain page (/domain). The Domain page renders directly from the domain's Intermediate Representation. It shows exactly the elements registered in the domain, so the diagram always matches the code. The page has three tabs, each showing a different view of the same model:

  • Topology draws an interactive, force-directed graph of your aggregates. Each node shows the aggregate name, its stream category, an ES/CQRS badge, and element counts. Directed edges connect aggregates linked by cross-aggregate event handlers and process managers. Zoom, pan, and drag nodes to arrange the graph. Hover a node to highlight everything it connects to, and use the mini-map to navigate large domains.
  • Event Flows shows the command-to-event-to-handler chains as a directed graph, grouped by cluster, with element-type filters for narrowing to just commands, events, or handlers.
  • Process Managers renders each process manager as a state machine alongside a summary of the domain's elements.

Click any aggregate to open the detail panel, a slide-in view with collapsible sections for its fields, entities, value objects, commands, events, handlers, invariants, and repository/models.

Browse and trace the timeline

Open the Timeline page (/timeline) to walk the event store chronologically. The filter bar narrows the stream by stream category, event type, command type, aggregate ID, and kind. The page loads more results as you scroll. Each row opens a detail panel with the full payload and metadata, plus links into the message's correlation chain and aggregate history.

Two sub-views drill in from there:

  • Correlation chain renders the causation tree for a correlation_id as a vertical graph, with handler attribution, per-step latencies, and cross-aggregate boundary markers. Deep-link straight to it with ?correlation=<id>.
  • Aggregate history lists one aggregate instance's events in order with version labels. Deep-link with ?stream=<category>&aggregate=<id>.

The Traces tab searches recent correlation chains by aggregate ID, event type, command type, or stream category, and links each result to its chain view. When the domain is actively processing, both the timeline and the correlation graph update live over SSE: new events animate in and the summary stat cards refresh without a reload.

Find a request from its correlation ID

Every HTTP response carries an X-Correlation-ID header. Paste that value into the Traces search (or ?correlation=<id>) to jump straight to the full causal chain for that request.

See also