Buckets
Buckets in the left sidebar is where you create buckets and work with their contents without writing code.
This page covers the console specifically. For the concepts, the S3 API equivalents, and code samples, see Buckets, Objects, Versioning, and Object Lock.
The bucket list
The table shows Name, Region (both the label and the region ID, such as Europe (France) and eu-west-1), and Created. Under each name a second line gives the bucket's storage used and object count, and an icon marks the bucket as private.
All of your buckets are listed together, sorted by name. Once you have 5 or more buckets, a Search buckets box appears above the table, the column headers become sortable, and a region filter appears when your buckets span more than one region. With no buckets yet the page reads No buckets yet and offers Create bucket.
If one region cannot be reached, an amber notice names it and the buckets in the other regions are still listed.
Each row has an actions menu with Browse objects, Copy bucket name, Copy S3 endpoint, and Delete bucket.
Deleting a bucket
Choose Delete bucket from the row's actions menu and confirm. Only empty buckets can be deleted: if the bucket still holds objects or object versions, the console says it is not empty and links to how to empty it. Use Empty bucket first, then delete. Deletion is permanent.
Delete bucket is only shown to Owners and Admins. It works in every region, including eu-west-1, where DeleteBucket is not available over the S3 API. See Buckets.
Creating a bucket
Click Create bucket and fill in:
- Bucket name. 3 to 63 characters, lowercase letters, numbers, and hyphens only, starting and ending with a letter or number. No dots, unlike AWS S3. Must be globally unique.
- Region. Europe (France)
eu-west-1, the default, or US East (Michigan)us-east-1. This sets the endpoint you connect to. - Object settings, three switches that nest inside one another, all off by default:
- Versioning keeps multiple versions of an object.
- Object Lock requires Versioning to be on, and stays greyed out until it is.
- Retention requires Object Lock, and stays greyed out until it is. Turning it on reveals a Default Retention Policy with Governance or Compliance mode (Governance by default) and a Lock period in Days or Years, defaulting to 15 days. The maximum is 100 years.
Turning a parent switch back off resets the ones nested inside it.
You can also expand Create new key to generate an access key alongside the bucket. It is pre-scoped to the bucket you are creating and to that bucket's region, and the submit button changes to Create bucket and API key. The secret is shown once, in a dialog. The console then opens the new bucket's page.
A panel headed Included by default confirms two things you do not configure: all data is encrypted at rest, and every bucket is private, so access requires an API key.
Name, region, versioning, Object Lock, and the default retention policy are all fixed when the bucket is created, in the console and over the API alike. Decide on them up front, and create a new bucket if you later need a different configuration.
Inside a bucket
Opening a bucket shows its region, storage used, and creation time, then three property cards: Versioning and Object Lock (each Enabled or Disabled), and Retention, which shows the default retention policy, such as Compliance · 15 days, or None. Encryption is on for every bucket, so it has no card of its own.
Once the bucket holds objects, Empty bucket and Upload object buttons appear in the header. While it is empty, the upload button sits in the empty state instead.
Two tabs follow:
- Objects (N) is the object browser.
- API Keys (N) lists the keys that can reach this bucket, plus an Access endpoints block with the copyable S3 Endpoint, S3 Path, and Region you need to configure a client. This tab is hidden for the Read-only role.
Browsing objects
The object table shows Name, Size, and Last Modified, plus Version and Status columns when versioning is enabled. Keys containing / are grouped into folder rows you can click into, with a breadcrumb to walk back out.
On a versioned bucket, a key with more than one version gets an expander and an n versions pill. Expanding shows the older versions, badged Latest for the current one and Delete marker where a delete marker sits on top.
Per-row icon buttons download or delete a single object. A checkbox on each row, and on the header, lets you select several at once. See Deleting objects.
The browser shows the first page of results, around 1,000 entries. When there are more, an amber Showing the first n notice appears, and selecting and deleting apply only to the objects listed. The count on the Objects (N) tab then comes from your bucket analytics, a daily reading that can lag recent changes, so expect it to be higher than the number of rows in view. Use the AWS CLI or an SDK to page through a full listing, or Empty bucket to remove everything. See Objects.
Uploading
Open a bucket and click Upload object. You can drag files or whole folders onto the drop zone, or use Select files and Select folder.
An optional Prefix field puts your files under a path such as images/. Two details are worth knowing:
- The prefix field appears when you select individual files. Files picked as a folder keep their own relative paths, which take the place of the prefix.
- The prefix starts empty each time rather than carrying over from the folder you were browsing, so set it on every upload where you want one.
Folder uploads skip system files that come along with a folder: .DS_Store, Thumbs.db, desktop.ini, anything whose path contains a segment starting with a dot, and __MACOSX folders. A Skipped n system files message tells you how many were left out. A file you select on its own is always uploaded, whatever its name.
Selecting files does not upload them. Click Upload n file(s) to start, and a per-run progress bar tracks the batch. If any file fails, a Retry n failed button retries just those rather than the whole set.
The console uploads each file in a single request, which the S3 single-PUT ceiling caps at 5 GB. For objects above that, use the AWS CLI or an SDK with multipart support. See Multipart Uploads.
The object detail page
Clicking an object shows its Name, Size, Bucket, S3 Path, ETag, and, where they apply, Version ID, Retention, Description, and Tags. Description and Tags are read from the object's description and tags user metadata, not from S3 object tagging, and the console upload flow does not set them. When retention is in Compliance mode a Compliance retention lock notice gives the date before which the object cannot be deleted.
A Version history card appears when the object has more than one version, listing each Version ID with its Status, Size, and Last Modified, and marking the one you are viewing. Clicking another row opens that version.
Version history is drawn from the object list you arrived through, so open the bucket first if you have followed a direct link to an object and the card is not showing.
To bring an earlier version back as the current one, download it and upload it again. See Versioning.
Downloading and sharing
Download is available on each row in the object list and in the object detail header. It works on any plan, including the trial.
Share object generates a link anyone can open without an API key. Choose an expiry of 1 hour, 24 hours (the default), 3 days, or 7 days, then Generate link.
Sharing is a paid-plan feature. During the trial the button is disabled, with a tooltip pointing to an upgrade.
Deleting objects
Click the trash icon on a row, or Delete object on the detail page, and confirm.
To delete several at once, tick their checkboxes. A bar shows n selected with Clear and Delete. Ticking a folder row selects everything under it, and the header checkbox selects everything at the current level.
The console always deletes the specific version on the row you chose. It does not write a delete marker. On a versioned bucket, deleting the row for the current version permanently removes that version, and the earlier versions are left as they are. To hide an object behind a delete marker instead, delete it over the S3 API without a version ID. See Versioning.
An object under an active retention period or legal hold is protected, and attempting to delete it returns that protection as an error. See Object Lock.
Emptying a bucket
Empty bucket on the bucket page removes every object in the bucket, including every earlier version and delete marker. Type the bucket name to confirm, then click Delete everything.
The deletion runs on the server, so it is not limited to the objects listed on screen, and it keeps running if you close the dialog. When it finishes, the dialog reports Bucket empty, Some objects remain (with how many could not be deleted, for example because they are under active retention), or Deletion stopped if the job could not complete.
Deleting objects, in bulk or one at a time, and emptying a bucket all need a role that can delete objects. Uploading and creating buckets need a role that can write objects and create buckets. Controls you cannot use are hidden, and opening the upload or create page directly explains that the action is not part of your role. See Roles and permissions.
While an account is past due or in a grace period, access is read-only and uploads and deletions resume once the payment succeeds. See Trial & Billing.