Skip to content

Set up storage connections ​

You can register and edit S3-compatible storage in Settings → Connections. You need Admin permission to manage connections.

For each connection, set the name, purpose, and allowed operations.
Connections screen with Research storage, Archive storage, and AWS S3 registered

For each connection, set the name, purpose, and allowed operations.

Register the endpoint and S3 credentials ​

Click Add and enter the name, endpoint, region, access key ID, and secret access key. To change a registered connection, click Edit in the list.

The endpoint in the image is an example.com sample. No real credentials are shown.
Screen for editing the endpoint, region, and credentials of an S3 connection

The endpoint in the image is an example.com sample. No real credentials are shown.

FieldWhat to set
NameDisplay name shown in the Connection selector in Storage and in the list
EndpointURL of the S3-compatible API
RegionRegion that the endpoint requires
Access key ID / Secret access keyS3 credentials issued for the endpoint
path-styleSet to match the endpoint's S3-compatible implementation
ListObjectsv1 or v2. Choose the one the endpoint supports

Credentials are encrypted when saved. If you don't want to change the existing credentials, leave the fields blank as the edit form instructs.

Choose the default connection for Storage ​

Click Set as default in the connection list to open that connection whenever you open Storage. You can switch to another connection from the Connection selector in Storage.

Limit which users can see a connection ​

By default, a connection is shown to all logged-in users. To show it only to some users, select Allowlist and specify the users to allow.

Users who are not allowed do not see the connection, and opening its URL directly returns 404. Admins who can manage connections still have access so that they can fix the settings.

Restrict allowed operations per connection ​

You can keep lists visible while restricting fetching file contents and writing back.
Screen for configuring preview, download, README editing, and other operations per connection

You can keep lists visible while restricting fetching file contents and writing back.

SettingWhen disabled
List buckets and objectsThe lists for this connection cannot be opened
File previews (text / image / audio / video)Text, images, audio, and video cannot be opened
File downloadsThe download action is not shown
Open archives (tar / tar.gz / tar.xz)The contents of tar and similar archives cannot be opened
Audio information and waveformsAudio analysis is not run
SpectrogramsSpectrograms are not shown
Read READMEsREADMEs are not shown (and cannot be edited)
Edit READMEsREADMEs cannot be written back to the bucket

The API also enforces these restrictions. Connection-level restrictions apply to Admins too, so these settings do more than hide buttons.

Configure scanning, scheduled measurements, and estimates ​

Allow directory scans and Allow bucket capacity scans are separate settings. To use scheduled measurements, enable them and specify the interval.

You can set the scan page size to 100, 250, 500, or 1,000 items. Reduce it for S3-compatible implementations that respond slowly to list requests.

Transfer estimates use each connection's read/write bandwidth, concurrency, provider, and capacity. Update these values once you have measured them.

Apache License 2.0