--- title: "Object Storage" description: "Create S3 compatible buckets and access keys, connect any S3 client, serve files through the CDN, and understand tiers, pricing and limits." category: "Deployment" language: en source: https://cubepath.com/docs/products/deployment/object-storage updated: 2026-11-17 --- # Object Storage Create S3 compatible buckets from **Object Storage** in [my.cubepath.com](https://my.cubepath.com). Any S3 client works with them: the AWS CLI, rclone, restic, boto3 and the other AWS SDKs, or `cubecli`. ``` https://eu.cubestorage.io (region eu) ``` > Note: Your account must be [verified](/docs/products/overview/account-setup) before you can create a bucket, and Object Storage is only available in projects billed **hourly**. ## What it's for Object Storage keeps files (objects) in buckets and serves them over HTTPS with the S3 API. It fits data you write through an application or a tool rather than a mounted disk: - **Backups and archives**: database dumps, server backups with restic or rclone, logs you need to keep. - **Application data**: user uploads, generated reports, exports. - **Static assets and media** served to the public through the [CDN](/docs/products/deployment/deploy-a-cdn). Buckets are **always private**. There are no public buckets and no static website hosting: to publish files, connect the bucket to a CDN zone (see **Serve files to the public** below). ## Tiers | Tier | Media | Best for | |------|-------|----------| | **Infrequent Access** | HDD | Backups, archives, logs and data that is written once and rarely read | | **Standard** | SSD | Coming soon. Content that is read often: media, CDN assets, application data | Every tier of a region shares the same endpoint. A bucket **cannot change tier** after it's created. ## Create a bucket ```steps ### Choose a storage tier Pick the tier that matches how often the data will be read. The card shows its prices and the free allowance. ![Storage tier selection with pricing for Infrequent Access](https://cdn.cubepath.com/img/docs/products/4adae0d9-4fd8-417a-9f21-5cb9a44e8c18.png) ### Name the bucket 3 to 63 lowercase letters, numbers and hyphens, starting and ending with a letter or number. Names are **unique across all CubePath customers**, so a common name may already be taken. ### Versioning (optional) Turn on **Enable versioning** to keep previous versions of overwritten or deleted objects. You can also turn it on later. ### Create Click **Create Bucket**. It's ready in 10 to 20 seconds. ``` > Note: Your first bucket may require a balance of at least **$5**. Each organization can have up to **20 buckets**. ## Create an access key S3 clients authenticate with an **access key**, not with your CubePath login or API token. Open **Object Storage → Access keys → Create access key**: - **Name**: what the key is for (e.g. `backups-db01`). - **Tier**: keys work on the buckets of one tier. - **Permission**: **Read and write**, or **Read only**. - **Bucket access**: **All buckets** of the project in that tier (including ones you create later), or **Only selected buckets**. - **Expiration** (optional): the key stops working at the end of that day (UTC) and is then deleted. > Warning: The **secret access key is shown only once**. Copy it or download one of the ready to use files (`.env`, `rclone.conf` or `~/.aws/credentials`) before closing the dialog. If you lose it, delete the key and create a new one. The key starts working a few seconds after it's created, once its status is **Active**. > Tip: Create one key per application, give it only the buckets it needs, and make it read only if it never writes. A leaked key then exposes one bucket instead of all of them. ## Connect an S3 client Use the endpoint `https://eu.cubestorage.io`, the region `eu` and your access key. **AWS CLI** ```bash aws configure set aws_access_key_id CPXXXXXXXXXXXXXXXXXX aws configure set aws_secret_access_key YOUR_SECRET aws configure set region eu aws s3 cp backup.tar.gz s3://my-bucket/ --endpoint-url https://eu.cubestorage.io aws s3 ls s3://my-bucket/ --endpoint-url https://eu.cubestorage.io ``` **rclone** (`~/.config/rclone/rclone.conf`) ```ini [cubepath] type = s3 provider = Other access_key_id = CPXXXXXXXXXXXXXXXXXX secret_access_key = YOUR_SECRET endpoint = https://eu.cubestorage.io region = eu ``` ```bash rclone copy ./backups cubepath:my-bucket/backups ``` **cubecli** ```bash cubecli s3 bucket create my-bucket --tier ia cubecli s3 key create --name backups --tier ia --bucket my-bucket --output rclone >> ~/.config/rclone/rclone.conf ``` Both URL styles work: `https://eu.cubestorage.io/my-bucket/file.txt` (path style) and `https://my-bucket.eu.cubestorage.io/file.txt` (virtual host). Uploads straight from a browser (CORS) only work with the **path style** URL. ## The bucket tabs | Tab | What it's for | |-----|---------------| | **Overview** | Size, object count, this month's usage and cost, and the connection details (endpoint, region, URLs) | | **Files** | Browse, upload, download and delete files, and create folders | | **CDN** | Serve the bucket's files to the public through a CDN zone | | **Settings** | Versioning, deletion protection and billing details | | **Activity** | Every change to the bucket and who made it | Size counts every stored version and refreshes about every 15 minutes. ### Files The **Files** tab is a file browser in the dashboard: open folders, **Upload** (or drag and drop onto the panel), **Download**, **Delete** files or whole folders, and create a **New folder**. Large files are uploaded in parts automatically. The browser works with your dashboard session; scripts and applications use an access key and any S3 client instead. Its requests are billed like any other request to the bucket. ### Settings - **Versioning** keeps every version of an object, so overwritten or deleted files can be recovered. Old versions are stored and billed like any other data. Once enabled, versioning can be **suspended** but never turned off. - **Deletion protection** stops the bucket from being deleted from the dashboard, the API or the CLI until you disable it. ## Serve files to the public Buckets stay private. To publish their files, add the bucket as an **origin** of a [CDN zone](/docs/products/deployment/deploy-a-cdn): open the zone → [Origins](/docs/products/deployment/cdn-origins) → **Add origin** and pick the bucket. The **CDN** tab of the bucket links you there and shows the public URL once it's connected. The CDN reads the bucket with a dedicated read only key that only works from the CDN edges, and caches the files close to your visitors. Your bucket's own access keys never leave your applications. - Traffic from the bucket to the CDN is **not billed as egress**: you pay the CDN zone instead. - Requests from the CDN edges count as **class B requests** of the bucket. - One CDN origin per bucket. To stop serving it, delete that origin from the zone. ## Pricing Billing is hourly in arrears, from the moment a bucket is active until it's deleted. Nothing is charged while it's being created. | Infrequent Access | Price | |-------------------|-------| | Storage | $0.008789 per GiB-month ($9 per TiB-month) | | Egress (to the Internet) | $0.009766 per GiB ($10 per TiB) | | Class A requests (writes and listings: PUT, POST, COPY, LIST) | $0.006 per 1,000 | | Class B requests (reads: GET and HEAD of an object) | $0.0006 per 1,000 | | Deletes, ListBuckets, OPTIONS | Free | **Free every month**, per organization and tier: **5 GiB-month** of storage, **5 GiB** of egress and **20,000 requests**. You get an email when the free allowance runs out. The **Usage** tab of **Object Storage** shows storage, egress and requests per bucket for the month, what has been billed so far and a projection for the full month. Failed requests (4xx and 5xx) are never billed. ## Limits - **20 buckets** per organization and **20 access keys** per project and tier. - **5 GiB** per single upload; larger files go in parts (multipart, up to 10,000 parts). The dashboard handles that for you. - **1 TiB per bucket** as a soft limit: you get an email when a bucket goes above it, and if it's still above after **72 hours**, uploads to that bucket are paused (reads, listings and deletes keep working) until it's back under. Contact support if a bucket needs more. - Presigned URLs last at most 24 hours. Unfinished multipart uploads are cleaned up after 7 days. ## Delete a bucket From **Settings → Delete bucket**, type the bucket name to confirm. - By default the bucket is only deleted if it's **empty**. Tick **Also delete every object in the bucket** to delete all objects, versions and unfinished uploads first. This can't be undone and takes a while for large buckets. - A protected bucket, or one that is still served by a CDN origin, can't be deleted until you disable protection or remove the origin. - Billing stops once the deletion finishes. The name stays **reserved for your organization for 90 days**. > Note: If a bucket shows as **suspended**, your data is kept but can't be reached. When it's because your organization is suspended, topping up your balance restores access.