Overview

This flow represents the golden path through the ironstar platform, from initial authentication through data analysis and dashboard creation. It captures the typical journey of a data analyst using the system.

Workflow stages

The workflow progresses through distinct phases.

Authentication establishes user identity via OAuth (GitHub or Google) and creates an authenticated session.

Catalog selection allows the user to choose which DuckLake catalog to work with, triggering metadata refresh to populate the dataset browser.

Query execution represents the core analytical work, where SQL queries are executed against the selected catalog. This phase has three possible outcomes: successful completion, failure (with error details), or cancellation by the user.

Dashboard creation enables visualization of query results through chart widgets organized into tabbed layouts.

Query persistence allows saving queries for reuse, completing the typical work session.

Branching logic

The flow includes decision points where execution may diverge.

Query execution branches into three outcomes based on whether the query completes successfully, encounters an error, or is cancelled by the user.

After adding charts, users may either organize them into tabs or proceed directly to saving queries.

Session lifecycle

Sessions are refreshed periodically while active. The session may terminate through explicit logout (invalidation) or TTL expiration. See the SessionLifecycle flow for detailed session state transitions.