Limitations

The following are a few limitations to consider while using IONOS Object Storage:

• Access keys: A user can have up to five access keys.

• Storage size: The minimum storage size available is 1 Byte of data and is extendable to a maximum storage of petabytes.

• Bucket naming conventions: Only buckets for static website hosting can use dots (.) in the bucket names. For more information, see bucket Naming conventions.

• Bucket count: A user can create up to 1000 contract-owned buckets and 500 user-owned buckets. For more information, see Bucket Types.

• Bucket Policy size: The maximum allowed Bucket Policy size for a contract-owned bucket is 1 MiB, and for a user-owned bucket is 20 KiB.

• Object size: The maximum allowed object size is 5 TiB.

• Object name length: The maximum allowed length of the folder path, including the file name, is 1024 characters.

• Maximum Number of Objects per bucket: The maximum number of objects per bucket is limited to 200 million objects per unversioned bucket and 50 million objects per versioned bucket. If you reach this limit, you may not be able to add further objects to the bucket. To continue using the IONOS Object Storage, please create a new bucket. Buckets that have reached the limit can still be accessed to read or delete objects. Please note that a bucket previously configured with versioning, but now set to unversioned, may not qualify for the higher quota.

• File upload size: A file upload size cannot exceed 5 GiB (5368709120 bytes) for contract-owned buckets and 4,65 GiB (5000000000 bytes) for user-owned buckets. If you have a single file exceeding this limit, you can bypass it using multi-part uploads. CLI tools such as AWS CLI and graphical tools such as Cyberduck automatically handle larger files by breaking them into parts during uploading.

• Bandwidth: Each connection is theoretically capped at approximately 10 G per region. However, remember that this is a shared environment. Based on our operational data, achieving peak loads up to 2x7 G is feasible by leveraging parallel connections. However, this is on a best-effort basis and without any guaranteed Service Level Agreement (SLA).