Search⌘KHide sidebarOpen menu
Switch to dark mode
Copy page content as Markdown⌘⌥C
Login
Query and datasets

The Data Landscape graph

Every connection, dataset and pipeline on one map.

Home is the Data Landscape graph, where you land after signing in. From here you create connections, pipelines and datasets and follow their runs. Each card is a node, and lines run left to right in the direction data flows.

Nodes and lines

NodeWhat it isClickDouble-click
Object storeA bucket registered in Settings. The default store is labeled Beetl storage.Opens Register data set on that store
Postgres sourceA database added with New connection, read in placeOpens Register data set on that source
ConnectionAn integration or webhook connectionOpens the details paneOpens its setup, where available
File setsOne card for all your File SetsOpens the File sets dialog
Data setAny dataset. Registered and external ones are labeled.Opens the details paneOpens Query with SELECT * … LIMIT 100
PipelineA pipeline, with its source and stage countsOpens the details paneOpens the editor

A line goes into a dataset from whatever writes it (a connection, the File sets card or a pipeline) or from the store or source it was registered from, and out of a dataset to each pipeline that reads it. A dataset Beetl writes has no line to the store holding it; its details pane names the store. Click or hover over a node to highlight its lineage upstream and downstream.

Status markers

MarkerOnMeaning
Running, with a spinnerPipelineThe latest run is queued or running.
SpinnerConnectionA sync is running or waiting to start.
Warning trianglePipeline or connectionThe pipeline's last run, or a sync's latest run, failed.
Warning triangleData setNothing has been written yet. Hover to see the status, usually Pending.
Check markPipelineDeployed, with no unpublished changes.
DotPipelineThe Working version has changes that are not deployed.
Status dotConnectionGreen when active, grey when paused or waiting for authentication. Hover to see which.

The details pane

Click a connection, dataset or pipeline to open its details pane on the right. Drag its left edge to resize it, and close it with × or by clicking empty canvas.

NodeTabs
Data setDetails: status, type, storage location, object store, partitioning, origin, schema and tags. Data preview: the first 30 rows, with Open in Query.
PipelineDetails: versions, deployment, sources, required run variables and the latest run. Runs.
ConnectionDetails: type, authentication, and each sync with its latest run. Runs and Logs.

You can filter runs by time range, and a connection's runs also by sync. On the Details tab you add and remove tags, and the buttons at the bottom repeat the node's right-click actions. See Runs.

Selected customers dataset on the Data Landscape with its Details pane showing Initialized status, storage information and a four-column schema
Select a dataset to inspect its status and schema. Switch to Data preview to see its rows.

Right-click menus

Right-click empty canvas to create something:

ItemWhat it does
New pipelineCreates an untitled pipeline and opens the editor.
New connectionOpens setup for an integration, a webhook or a Postgres source. See Connect your data.
Register data setRegisters files in an object store or a Postgres table. See Registered datasets.
Add zoneAdds a zone where you clicked.

Right-click a node to act on it:

NodeActions
PipelineOpen editor, Run pipeline (deployed and enabled only), Disable or Enable, Delete
ConnectionConfigure, Pause or Resume, Delete
Data setQuery (once data has been written), Delete
Object storeBrowse object store
Postgres sourceRegister table, and Delete for a TenantAdmin
ZoneDelete zone
Several selected nodesDelete all of them

Every delete asks you to confirm. Deleting a connection keeps its datasets, and deleting a dataset keeps its data in storage. See Connection types, connections and syncs.

The + on a card

Click the + on a dataset card's right edge, enter a pipeline name and choose Create pipeline. The pipeline appears on the canvas with that dataset as a source called source. On an integration connection, the + lists the integration's streams and opens the sync dialog with your choice filled in. See Integrations.

Search and tags

Find anything… at the top left dims every node whose name does not match. The Tags menu keeps only nodes carrying a tag you check. Only connections, datasets and pipelines take tags, so other nodes dim while a tag is checked.

Arrange the canvas

Drag across empty canvas to box-select, Cmd/Ctrl-click to add or remove a node, and nudge the selection with the arrow keys. Delete or Backspace deletes it once you confirm. In the toolbar, the magnet turns on Snap to grid, Tidy layout arranges cards by dependency, and ? lists these shortcuts. The bottom right holds a clickable minimap, zoom, fit to view, and undo and redo for navigation.

A zone is a colored background for grouping nodes. Double-click its label to rename it, select it to resize it or change its color, and drag it by its border. Moving a zone leaves its nodes behind, so box-select both to move them together. Tidy layout resizes zones to fit their nodes.

Dropping files onto the canvas turns each file into its own File Set.

The layout is saved in your browser

Node positions, zones and the snap setting are stored in this browser, per tenant. Other users, and you on another device, see their own arrangement, and clearing site data resets it.

The assistant on Home

Open Assistant from the right edge of the window. On Home it sits below the details pane, with a divider you can drag.

The assistant reads the graph and can arrange it: move nodes, add or resize zones, set the search and tag filter, and select a node. These edits never change the connections, datasets or pipelines themselves. Its reply ends with Draft updated, where Dismiss undoes that turn. See The agent trust boundary.

Assistant selecting the Consolidated ERP Invoices Pipeline on Home, highlighting its three source datasets and opening its details pane
Ask the assistant to find a pipeline on Home. Its sources are highlighted, its details open above the conversation, and Dismiss undoes the selection change.

Pages not in the sidebar

These three list pages have no sidebar entry. Open them by address:

PageAddressAlso reached from
Connections/web/connectionsClosing a connection setup opened as a full page
Data Catalog/web/data-sets
Pipelines/web/pipelinesThe Pipelines link in the full-page pipeline editor

Last updated on

On this page