Exporting your data

What is in the archive

Everything you have in the workspace, in one zip: every job, the whole run history including the outputs not yet swept, every account, and the audit log itself. A zip and not a tar, because it is written row by row and need not know its size in advance — a run history can be large, and an export must not depend on how long you have been a customer.

The zip does not contain the values of your secrets — Runjet never had them — nor the agents' keys.

The procedure

The Export screen, for the admin role: Prepare an archive, a Building the archive page refreshing itself, and then The archive is ready with a link to download. The build runs outside the request, so a large history does not run into the browser's timeout. You can ask again at any time (Prepare another archive); every request is a workspace.export row in the log — an archive outlives the workspace, and it is worth knowing who took one.

An hour, and once

The link lasts an hour and works for one completed download; once the archive has gone out whole, the screen says That link has been used. An archive is as sensitive as the account it came from — it holds whatever your commands printed into their logs — and a forwarded link must not be something that still works. Asking again is one click.

A download that breaks off — a dropped connection, a laptop closed half-way — does not use the link up: click Download again while the hour lasts. While one download is in progress a second one is refused, so two copies cannot leave through the same link at once.

From a suspended workspace too

An export is a read, and reads work in a suspended workspace. That is a requirement rather than an accident: the last warning email says "take your data while it is still readable" twenty-eight days after scheduling stopped, and it has to be possible.