Documentation by ClonePartner

Start typing to search.

Using Envoy

Data Explorer

View Markdown

Browse connector resources, build visual or raw queries, export results, and save reusable team presets with parameters.

Data Explorer

Data Explorer is the interactive query surface for stored Connectors. Operators use it to browse tables, collections, files, and API resources, preview rows, run bounded queries, and save useful queries as team presets — without writing a task.

Screen overview

Open Data Explorer from the sidebar. The screen has two panels:

  • Resource sidebar (left) — a connector picker on top, then the list of resources the connector exposes: tables, collections, files, folders, or API resources. A Filter... input appears for longer lists, and a refresh button reloads the schema, bypassing the server's connector cache. For connectors that support uploads (CSV, YAML, S3, Azure Blob), an Upload file button sits above the list.
  • Main panel (right) — the query controls for the selected resource, then the results.

Before a connector is chosen the main panel shows "Explore your data — Choose a connector to browse tables, preview rows, and run queries." After choosing one, pick a resource from the sidebar or the Browse resources… picker, which has a Search resources… field.

Once a resource is selected, the controls bar shows the connector badge, a searchable resource picker, a Presets menu, an Upload button (file-capable connectors only), a row-limit picker (labeled "N rows"), and the primary action button — Run for most queries, or Count, Insert, Update, Delete to match the selected action. While a query runs, the button becomes Stop.

Visual and raw modes

SQL, MongoDB, and Dynamics 365 connectors show a query builder with two tabs, Visual and Raw, plus an action toggle:

  • Find and Count are available on all builder types.
  • Insert, Update, and Delete appear only for MongoDB, and only for roles allowed to mutate. Update and delete require a non-empty filter, and both show a confirmation dialog with the count of matching documents before writing.

Visual mode

Visual mode builds the query from structured controls:

  • Add filter adds a row with a field picker, an operator, and a value. MongoDB filters include a value-type selector so ObjectId, date, number, boolean, and null values match their stored BSON types.
  • Add sort adds server-side sort rows.
  • The row-limit picker caps returned rows (preset options plus a Custom limit… input).
  • The column picker (with Search columns..., Select all, and Unselect all) projects specific columns.

Raw mode

The Raw tab opens a code editor for connector-native queries: SQL for database connectors ("Enter a SELECT query"), JSON with a method key for MongoDB (find, aggregate, count, insert, update, delete), an OData $filter expression or JSON for Dynamics 365, and JSON method calls for S3. Press Ctrl+Enter (Cmd+Enter on macOS) to run.

The server validates every explorer query: only methods a connector declares read-only (plus role-gated write methods) are allowed, SQL is restricted to SELECT, WITH, and EXPLAIN statements, and MongoDB aggregate pipelines reject $out and $merge stages. Mongo write methods require an admin or superadmin role and are audited.

Results

Results render in a Table or JSON tab, with the loaded row count and a truncation note when more rows matched. Hovering a row reveals a copy action; double-clicking a cell copies its value. A JSONata panel applies a display-side transform to the loaded rows.

The download menu exports data two ways:

  • Export CSV / Export NDJSON — the currently loaded rows (JSONata transform applied).
  • Export all as CSV / Export all as NDJSON — every row matching the query, streamed from the server without a row cap. Available for read-only queries only; the JSONata filter is not applied. A running export can be canceled from the same menu.

Shareable URLs

The explorer keeps its full state in the URL: the connector ID in the path (/data-explorer/42) and the resource, mode, action, filters, columns, sort, limit, raw query, result view, and JSONata expression in query parameters. Copy the browser URL to share an exact view with a teammate, or bookmark it to return later. Opening a shared URL restores the query and re-runs it when it is a read query — URLs containing raw Mongo write payloads or Insert/Update/Delete actions never auto-run.

Presets

Presets are saved, team-shared queries. The Presets menu on the controls bar lists Built-in and Team presets with a Search presets… field; selecting one restores its saved state.

To save a preset

  1. Build the query you want to share.
  2. Open Presets and select Save current as preset…. The Save team preset dialog opens.
  3. Enter a Name. The dialog notes that it "Saves current filters, columns, mode, limit, and JSONata for this connector and collection. Visible to all team members."
  4. Select Save.

Give presets purpose-based names and avoid embedding personal or secret values — every team member sees them.

Preset parameters

Parameters turn parts of a query into prompts, so the same preset serves values that change on every run (a job ID, a tenant, a date):

  • In the Save team preset dialog, select Make parameter next to a visual filter to promote its value to a parameter.
  • In raw mode, place {{name}} placeholders inside the JSON query text; the dialog detects them and lists the resulting parameters. Parameterized raw presets require a query that parses as JSON — SQL and OData raw queries cannot be parameterized.

Parameters have three types, validated when the preset is saved: string (free text), objectId (a 24-character hex MongoDB ObjectId), and mongoRecord (a searchable record picker over a Mongo collection on the same connector). Every placeholder must reference a declared parameter and every declared parameter must be used, or the save fails with a validation error.

To run a preset

  1. Open Presets and select the preset. Parameterized presets show a parameter count next to their name.
  2. If the preset has parameters, a dialog prompts for each value — free text, an ObjectId, or a searchable record picker depending on the parameter type.
  3. Select Apply & run. The explorer substitutes the values and executes the query.

To manage presets

Open Presets → Manage team presets…. The Team presets sheet lists all team presets with search, apply, rename, and delete actions. Anyone can create presets and edit their own; deleting requires an admin or superadmin role. A preset is scoped to the connector and resource it was saved on; presets saved without that scope appear everywhere.

Truto and catalog connectors

Truto and catalog-backed connectors use a request builder instead of the filter tabs. Pick an integrated account in the sidebar (catalog connectors show their single pinned account), then choose a resource from the All / Unified / Proxy surface toggle. The builder has a Method select (list, get, create, update, delete, or Custom API), a Resource ID field for get/update/delete, a Response format select, Params and Body tabs (with an Edit as JSON toggle), and a request URL preview. Send executes the request; DELETE requests show an explicit confirmation. Write methods and row edit/delete actions appear only for roles allowed to mutate.

File and object storage connectors

CSV, YAML, S3, and Azure Blob connectors browse files rather than tables. S3 browsing is hierarchical with breadcrumbs and paged listing (Load more). Selecting an object shows its type, size, and modified date with a Download button; text and image objects preview inline, while binary files stay download-only. The Upload file button opens an upload dialog scoped to the connector's accepted extensions and the server's upload size limit.