Workspace

UI overview

After signing in, you land on the Home view: recent assets, failing syncs, and Company Brain entry. The left rail switches between Catalog, Lineage, Editors, Metrics, Agents, and Settings. A global command palette (Ctrl+K) jumps to any asset, action, or doc page.

The top bar shows active environment (production vs staging), last sync timestamp per connector, and your role. RBAC controls which databases, metrics, and agent tools appear: platform admins see connector credentials; analysts may see certified metrics only.

Catalog

The catalog is faceted search across all graph nodes: tables, dbt models, DAGs, topics, dashboards, glossary terms, and people (owners). Filters include asset type, connector source, PII tag, certification status, and freshness.

Each asset card shows description, owner, last sync time, upstream/downstream counts, and quick actions: open lineage, ask Brain, copy URI. Column-level detail appears when connectors expose it; dbt and Snowflake typically populate full column lists with types and tags.

Asset detail panel

Clicking an asset opens a split view: metadata and tags on the left, mini-lineage on the right. Tabs cover Schema, Lineage, Quality (dbt tests), Discussion (threaded comments), and History (definition changes, certification events).

Use the catalog to answer "who owns this table?" and "what tags apply?" without querying INFORMATION_SCHEMA. Stewardship workflows let owners claim unowned assets and request certification reviews.

Lineage graph

The lineage view renders an interactive DAG of dependencies. Pan and zoom; click nodes to expand upstream or downstream. Column-level mode highlights individual field paths when metadata supports it, essential for GDPR impact analysis and rename planning.

Blast radius

Select a node and press B or choose Preview impact. Metroflow highlights all downstream dashboards, exports, ML features, and tasks within N hops. Export the subgraph as PNG or structured JSON for PR attachments.

Cross-layer traces

Follow an Airflow task through a dbt model to a Snowflake table and into a Looker explore in one continuous path. Environment toggle re-renders the graph from staging or production connector snapshots.

Compare mode (Enterprise) overlays two graph versions to see what changed between syncs, useful after large dbt refactors.

Editors

Context-aware editors bring stack knowledge into where you write SQL and YAML. The dbt editor validates ref() and source() against live manifests; stale refs show squiggles before CI runs. Inline lineage preview opens beside the editor without leaving the file.

SQL workspace

Ad-hoc SQL against your warehouse with autocomplete from synced schemas. Queries are local to your session; Metroflow does not persist result sets. Save snippets as linked assets for team reuse.

Manifest & config editing

Edit schema.yml descriptions and tests with templates from the document agent. Proposed changes open as diffs for review; merge through your normal Git flow; Metroflow does not push directly to main.

Metrics registry

The metrics registry lists certified and draft KPIs: MRR, churn, active users, funnel conversion, and custom definitions. Each metric page shows canonical SQL or dbt model link, steward, approval date, reconciliation status, and dependent dashboards.

Stewards submit definitions for certification; reviewers compare upstream lineage and collision with existing metrics. Certified metrics appear in catalog search with a badge and are preferred by Company Brain when interpreting business questions.

Reconciliation tab plots variance between declared sources (e.g. finance ledger vs product analytics) when validation queries are configured. Discrepancies link to the offending pipeline nodes.

First-week workflow for data engineers

A practical onboarding sequence once connectors are syncing:

1. Day 1: Verify the graph

   Confirm dbt model count and Snowflake table count match expectations. Fix connector filters or credentials before inviting the team.

2. Day 2: Claim ownership

   Search for assets in your domain; assign yourself owner. Add descriptions to the ten most queried tables; document agent can draft from lineage context.

3. Day 3: Trace critical paths

   Pick revenue or ops KPIs; walk lineage from dashboard to raw source. Note gaps for connector tuning or dbt exposures.

4. Day 4: Certify one metric

   Work with analytics lead to certify a single KPI (e.g. weekly active users). Link dbt model, tests, and downstream dashboards.

5. Day 5: Blast-radius drill

   Simulate renaming a column on a staging model; export impact subgraph. Share in team channel to validate graph completeness.

By end of week one, your team should trust search and lineage for daily debugging, reducing Slack "where does this field come from?" threads.

Keyboard shortcuts

Shortcuts work when focus is not inside a text field. Press ? in the workspace for the full cheat sheet.

Live demo

Explore a pre-populated workspace without self-hosting. The public demo includes sample dbt, Airflow, and warehouse lineage, Company Brain queries, and lineage blast-radius, ideal for evaluating UX before deploy.

Open the live demo →