S3.

Where S3 fits on a long engagement.

S3 is durable and cheap enough that most estates accumulate data in it indefinitely without deciding to. The cost per gigabyte is low, which is exactly why the total becomes significant: nobody notices a bucket growing until it is a line item.

Lifecycle policies are the answer and are routinely absent. Data that is accessed constantly for a month and rarely afterwards belongs in a cheaper class after that month, and data with no retention requirement belongs deleted. Both are policy decisions that cost nothing to implement and are simply never made.

What an assigned team does with S3.

Public bucket exposure remains one of the most common serious cloud findings, despite years of tooling to prevent it. Block Public Access at the account level is the control that actually holds, because it cannot be overridden by a bucket policy written in a hurry.

Verifying those controls across an estate is part of a security baseline, assigned under devops managed services.

What we use S3 for.

• Storage cost brought under control Lifecycle policies moving or expiring data, which is usually a large saving nobody has claimed.
• Public exposure prevented structurally Account-level blocks that a hurried bucket policy cannot override.
• Versioning where data matters Object versions retained, so an overwrite or delete is recoverable.

How S3 capacity is assigned.

Storage work is assigned inside platform capacity, with lifecycle and access policy set deliberately rather than left at defaults.