--- cuppy_hub: gcp cuppy_path: reference/storage/cloud-storage.md cuppy_section: docs cuppy_doc: 5EJY-KT49 cuppy_doc_uuid: d9574a0f-fcb3-42d0-bd5a-bb457a3eb412 cuppy_permalink: https://cuppy.io/gcp/d/5EJY-KT49 cuppy_version: 1 cuppy_updated: 2026-10-23T16:16:52.575Z cuppy_canonical: https://cuppy.io/gcp/docs/reference/storage/cloud-storage cuppy_mcp: https://cuppy.io/mcp cuppy_body_sha256: b76ede08b5727fd340429fcba18cfc25dcb03a4d53dc9d1091187e0b70deca85 # cuppy_body_sha256 is over the document body ALONE, excluding both generated blocks. --- # Cloud Storage — classes and locations ## 1. The five classes | Class | API name | Minimum storage duration | Retrieval fee | Stated availability | |---|---|---|---|---| | Rapid | `RAPID` | none | no | >99.95% (zones) | | Standard | `STANDARD` | none | no | >99.99% multi/dual-region; 99.99% region | | Nearline | `NEARLINE` | **30 days** | yes | 99.95% multi/dual; 99.9% region | | Coldline | `COLDLINE` | **90 days** | yes | 99.95% multi/dual; 99.9% region | | Archive | `ARCHIVE` | **365 days** | yes | 99.95% multi/dual; 99.9% region | `STANDARD` is what you get if you specify nothing. **The 365 days is the number to carry into any estimate.** It is Cloud Storage's longest commitment and it is twice the equivalent commitment on either of the other two major clouds — see `guides/parity/storage.md` §2, which is the one parity row most likely to change an archive design. ## 2. Minimum storage duration is a billing commitment, not a retention lock The object can be deleted whenever you like. The *charge* is the commitment. So a lifecycle rule that moves objects to `ARCHIVE` and a retention policy that deletes them after six months are, together, a bill for twelve months of archive storage on data you no longer have. **Check the lifecycle rule against the class's minimum duration before you ship it**, in both directions. ## 3. Four location types, and one that is class-restricted | Location type | What it means | |---|---| | **zone** | zonal — **`RAPID` storage only** | | **region** | one region | | **dual-region** | two geographically separate locations | | **multi-region** | geographically dispersed, worldwide | Two readings: - `RAPID` is the only class that can be zonal, so "zonal object storage" and "a cheap archive" are not combinable choices; - dual-region and multi-region are **location types of the bucket**, not replication you configure afterwards. A bucket's location is fixed at creation in the model this page describes, which makes it one more create-time decision to get right (the same family as Cloud KMS's, see `reference/identity/secret-manager-and-kms.md` §2). ## 4. The class is per object, and a bucket default is only a default > "A storage class is a piece of metadata that is used by every object" And: **changing a bucket's default storage class does not affect existing objects.** This is the fact that makes "we switched the bucket to Coldline" a false economy claim — it applies to new objects only. Existing objects move when a lifecycle rule or an explicit rewrite moves them, and each move is itself a billable operation. ## 5. What this document does not say No prices, and no adjustable quota values — `reference/overview/quotas-and-limits-model.md` explains why and names the call. The retrieval-fee column records **that a fee exists**, which is the design-bearing fact; the amount is a live value. Next: `reference/storage/cloud-storage-access-control.md` (the single most consequential setting on a bucket), `reference/storage/errors.md` (the JSON API reason strings), `guides/parity/storage.md`. --- You are reading one file. The same content is free over MCP at https://cuppy.io/mcp with no signup: ranked search across all 91 documents in this hub, the merged view when you work in a layer, known workarounds matched to the error you hit, and depth control so you spend tokens on what you need. Start with the `_hello` tool. How to connect: https://cuppy.io/gcp/ Something wrong in this document? Report it with the `session feedback` tool on the same endpoint.