http:// instead of https:// in the S3 API endpoint.
Plain HTTP access can reduce encryption-related performance overhead in workloads that send many S3 API requests or use clients where TLS certificate configuration is resource-intensive. Use it only when the security tradeoff is acceptable, as HTTP doesn’t encrypt object data, request metadata or request headers in transit.
How HTTP access works
HTTP access is a bucket-level setting and is disabled by default. Contact support if you want to enable it for a specific bucket. When you request HTTP access, provide the bucket ID or name, the bucket region and one of the following access modes:- Disabled: HTTP access is not allowed. This is the default mode.
- Region-local: HTTP access is allowed only from Nebius AI Cloud virtual machines (VMs) in the same region as the bucket. Requests from other regions and from outside Nebius AI Cloud are rejected.
- All: HTTP access is allowed from any source, including VMs in any region and clients outside Nebius AI Cloud.
How to check the bucket HTTP access mode
How to check the bucket HTTP access mode
- Web console
- CLI
- Go to Storage → Object Storage.
- On the Buckets tab, check the bucket label. Buckets with HTTP access enabled show HTTP - Region or HTTP - All.
https:// with http://.
For example, if your HTTPS endpoint ishttps://storage.eu-north1.nebius.cloud, the HTTP endpoint ishttp://storage.eu-north1.nebius.cloud.
Authentication
Plain HTTP access supports only access key pair authentication.Supported requests
Plain HTTP access is supported for S3 API requests that work with objects, such asGetObject, PutObject, DeleteObject or ListObjects.
The following bucket information requests are also supported over HTTP:
HeadBucketGetBucketLocationGetBucketVersioning
Limitations
- Chunked uploads are not supported over HTTP. Use HTTPS for these uploads.
- HTTP access doesn’t change Object Storage encryption at rest. Stored objects remain encrypted by default.