Skip to content

Approve And Manage Requests

Every data access request needs a decision before it can be activated. This guide walks through reviewing, approving, denying, and revoking requests entirely from the dashboard.

There’s no separate admin queue — admins and requesters share the same Data Requests list and the same request detail pages. The difference is what you see there: an organization admin’s list shows every request across the organization, while everyone else only sees their own.

A Pending approvals notification in the header (the bell icon) gives admins a shortcut: it lists the 5 most recently submitted pending requests, each showing its title and who requested it. Click one to jump straight to its detail page, or click See Pending at the bottom to open the full list filtered to pending requests.

On the Data Requests page:

  • Six summary cards along the top — Open, Pending, Approved, Active, Denied, Revoked — each showing a count and acting as a shortcut filter when clicked.
  • A Status dropdown filtering the same way (All Status, Pending, Approved, Active, Denied, Revoked, plus broader Open/Closed groupings).
  • A search box that filters by title, description, requester name or email, and template name.
  • A Show closed requests link at the bottom of the default view, when there are denied, revoked, or expired requests to see.

There’s no way to filter or sort by risk level here — if you specifically need to work through high-risk requests, open each one to check its risk badge, or rely on the pending-approvals list (most recent first).

A request’s detail page shows everything relevant to the decision: its risk badge (low, medium, or high, inherited from its template or set when it was created), its title, description, and business justification, and — read-only at this point — the datasource(s), datasets, and columns it’s scoped to (see Configure Data Sources). If it came from a template, a Template badge names it. See Create A Data Access Request From A Template for how a requester builds and submits one in the first place.

For a table loaded from an SSH datasource’s files (see Loading Files Into A Request), the dataset’s row shows its number of files and their total size, with any warnings as badges. Expanding it shows its Dates (or All files when the whole schema loads) and a Files list grouped by day (the latest 31 days). The files are fixed when the request is created: approving it approves exactly that list.

If a table’s files go over the deployment’s warning limits (100 MB or 31 days by default; see Persistent Platform Configuration), a Large data load notice at the top of the page lists each warning by dataset. The request can still be approved; the notice is there so you can ask for a narrower date range first. Requests over the hard size limit can’t be created at all.

A dataset also carries a warning when the SSH datasource has skipped files in the dates the requester chose whose schema isn’t accepted or that are too large to check (see Skipped Files); files that could not be read aren’t counted in it. It reads, for example, 3 files in these dates are left out: their schema is not accepted on the datasource. The notice’s title shows which warnings the request carries: Large data load for size and day-span warnings only, Files left out for left-out files only, and Warnings for both. Those files aren’t part of the request and are never loaded. The request can still be approved. If the requester needs them, an administrator accepts their schema and the requester creates a new request (see Files That Aren’t Loaded).

On a pending request’s detail page, click Approve. A confirmation dialog asks:

This will approve the request and make it eligible for activation. Are you sure?

Confirm with Approve. There’s no field to add approval notes — the decision itself is all that’s recorded. Once approved, the status badge changes to approved, and the request is ready to be activated.

On the same pending request, click Deny instead. The confirmation dialog reads:

This will deny the data access request. The reason is shown to the requester on the request page.

The dialog has a Reason (optional) field. What you type there is stored with the request and shown to the requester as Denial reason on the request page, so use it to say what would make the request acceptable. Confirm with Deny. Once denied, the status badge changes to denied, and the page says the request was denied. The request itself can’t be reopened; the requester can start a new one from it with Request again.

Once a request is active, either the requester or an admin can end it before it expires on its own. Open the request — or use the ⋮ menu on its row from the Data Requests list — and click Revoke.

The confirmation dialog spells out the consequence:

This will terminate the active data environment for this request. Any active tunnel will be torn down with it.

Type revoke into the confirmation field — the button stays disabled until it matches exactly — then confirm. There’s no separate “suspend”: revoke is the only way to end an active request early, and it’s final for that request. The environment and any tunnel are torn down immediately, and the status badge changes to revoked.

A request that can no longer be explored has a notice at the top of its page that names what ended it:

  • Denied: “This data access request was denied”. The reason, when the approver gave one, is under Denial reason.
  • Revoked: “This data access request was revoked”, with the reason when one was recorded (for example datasource <name> was disabled).
  • Expired: “This data access request has expired”, with the date its validity ended. A request the platform revoked because its validity ended says so.

Explore data is disabled on all three.

A denied, revoked or expired request has a Request again action: a button on the request’s page, and an entry in the ⋮ menu on its row in the Data Requests list (use Show closed requests to see those rows). It never changes the original.

  • For a request that was built by hand, it opens Create Data Access Request filled in with the original’s name, description, justification, datasources, tables, fields, filters, joins and dates. Change what you need, such as the scope an approver asked you to narrow, and submit. The new request goes through approval like any other.
  • For a request that came from a template, it opens the template, as if you had chosen it from the list. Fill in the details and parameters again.

The form is filled from the catalog as it is now, so it tells you about anything it couldn’t copy:

  • A table that is no longer in the datasource is left out, with a note naming it.
  • A field the original had that the catalog no longer has is not ticked, with a note naming it.
  • A join or row filter that can’t be copied exactly is called out. Without it the request would load more rows, so add it again before you submit.
  • A datasource that an administrator still has to update is left out.

An SSH table that an administrator renamed since the original request is found under its new name, with nothing to fix.

Requests created before this feature was installed kept no record of what was picked, so Request again can’t copy them: build the new request from the start.