canonical: https://jentic.com/apis/googleapis.com/filestore

# Google Cloud Filestore API

The Cloud Filestore API provisions and manages fully-managed NFS file servers on Google Cloud, providing shared persistent storage for Compute Engine and Google Kubernetes Engine workloads. It exposes operations to create instances in specific regions, take backups, restore from backup, promote replicas, and list available locations. Use it to automate the lifecycle of shared NAS-style file systems without provisioning storage hardware.

## For AI agents

Provision and manage Google Cloud Filestore NFS file server instances, backups, and replicas so an agent can automate shared file storage for cloud workloads.

## Scope

Does not store object blobs, serve HTTP file downloads, or back mobile clients directly - use for managed NFS file server provisioning and backup only.

## Capabilities

- Provision a managed NFS file server in a chosen Google Cloud region
- Snapshot a Filestore instance into a backup for disaster recovery
- Restore a Filestore instance from a previously created backup
- Promote a replica Filestore instance to a standalone primary
- List available Filestore locations and ongoing long-running operations

## Use cases

### Shared NFS Storage for GKE Workloads

Provision a managed Filestore instance to give multiple Google Kubernetes Engine pods read-write access to a shared file system without operating an NFS server. The API creates the instance in a chosen tier and region, exposes the mount address, and lets automation poll the long-running operation until ready. Integration into a Terraform-style pipeline takes a few hours.

Example prompt: Call POST /v1/{+parent}/instances to create a BASIC_HDD Filestore instance named shared-data in us-central1-a, then poll the returned operation via GET /v1/{+name}/operations until it reports done.

### Scheduled Backups and Disaster Recovery

Create on-demand backups of a Filestore instance via POST /v1/{+parent}/backups, then restore them with POST /v1/{+name}:restore when data corruption or accidental deletion occurs. Backups are managed objects discoverable through GET /v1/{+parent}/backups, supporting compliance and recovery-point-objective requirements without external snapshot tooling.

Example prompt: Trigger POST /v1/{+parent}/backups for instance shared-data, wait for the operation to complete, and tag the resulting backup resource with description nightly-2026-06-10.

### Replica Promotion for Regional Failover

When a primary Filestore instance becomes unavailable in a region, the API can promote a replica to a standalone instance via POST /v1/{+name}:promoteReplica, restoring write access for downstream workloads. This automates the failover step that would otherwise require manual coordination. Promotion is irreversible and best suited to genuine failover scenarios.

Example prompt: Call POST /v1/{+name}:promoteReplica on the replica resource and verify success by polling GET /v1/{+name} until the instance state reads READY without a sourceReplica field.

### Agent-Managed Filestore Lifecycle via Jentic

An AI agent connected through Jentic can manage Filestore instances end to end - provision, back up, restore, and clean up - using natural-language intents. Jentic exposes the 15 Filestore endpoints, handles OAuth 2.0 authentication transparently, and lets the agent string together create, poll, and snapshot calls without dealing with raw REST URLs.

Example prompt: Through Jentic, search for provision a managed nfs file server, load the create instances operation, and execute it with project=my-project, location=us-central1-a, tier=BASIC_HDD.

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | /v1/{+name}:restore | Restore a Filestore instance from a backup |
| POST | /v1/{+name}:revert | Revert a Filestore instance to an earlier snapshot |
| POST | /v1/{+name}:promoteReplica | Promote a replica instance to standalone primary |
| GET | /v1/{+parent}/backups | List backups under a project location |
| GET | /v1/{+name}/locations | List Filestore locations |
| POST | /v1/{+name}:cancel | Cancel a long-running Filestore operation |

## Key resources

- **Instances** — Create, get, patch, delete, restore, revert, and promote managed NFS file server instances
- **Backups** — Create, list, get, patch, and delete backup snapshots of Filestore instances
- **Locations** — List the regions and zones in which Filestore is available
- **Operations** — Track and cancel long-running operations spawned by instance and backup actions

## Why Jentic

- **Setup:** Wiring the Cloud Filestore API by hand means standing up Google OAuth 2.0 with a service account, picking the right location for each managed NFS instance, and polling the long-running operations yourself. Through Jentic you install once, import the Cloud Filestore API from the API Directory, store the service-account credential once, and your agent calls it.
- **Permission scoping:** Filestore puts the instance and backup resource in the URL path (/v1/{name}:restore, /v1/{parent}/backups), so a rule can pin your agent to one instance or backup and nothing else. You choose the operations it may call, so destructive ones like revert or promoteReplica are not included unless you add them.
- **Credential handling:** Your Google service-account credential is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'provision a managed NFS file server' or 'restore a filestore backup', and Jentic returns the matching Cloud Filestore operation with its input schema so the agent calls the right endpoint without browsing the reference docs.

## Related APIs

- **Google Cloud Storage API** — Object storage for unstructured blobs rather than POSIX-compliant shared NFS.
- **Firebase Storage API** — Mobile and web client-facing object storage backed by Cloud Storage.
- **Google Cloud Billing API** — Tracks the spend generated by provisioned Filestore instances and backups.

## FAQ

### What authentication does the Cloud Filestore API use?

The API uses Google OAuth 2.0 with the cloud-platform scope. Through Jentic the OAuth credentials live in your Jentic One instance and a short-lived access token is supplied to each request without entering the agent's context.

### Can I create and restore Filestore backups with the Cloud Filestore API?

Yes. POST /v1/{+parent}/backups creates a backup of an instance and POST /v1/{+name}:restore restores an instance from a named backup. Both operations are long-running and return an operation resource you can poll under /v1/{+name}/operations.

### What are the rate limits for the Cloud Filestore API?

Google Cloud enforces standard project-level quotas for write and read operations on Filestore. Concurrent instance creation is also bounded by per-region capacity quota, which can be raised in the Google Cloud console quota page.

### How do I provision a Filestore NFS instance through Jentic?

Run pip install jentic, then search for provision a managed nfs file server, load the create instances operation, and execute it with project, location, instanceId, and tier parameters. Get started with Jentic One, the self-hosted execution layer.

### Does the Cloud Filestore API support promoting replicas?

Yes. POST /v1/{+name}:promoteReplica promotes a replica Filestore instance to a standalone primary, used during regional failover scenarios. The action is irreversible once the operation reports done.

### Is the Cloud Filestore API free?

The API calls themselves are not separately billed, but the Filestore instances and backups they create are charged by tier, capacity, and region under standard Filestore pricing in the Google Cloud price list.

### Can I limit what my agent is allowed to do with the Cloud Filestore API?

Yes. Because you run Jentic One yourself, your own rules decide which Cloud Filestore operations the agent may call and which credential it uses. Filestore puts the instance and backup resource in the URL path, such as /v1/{name}:restore and /v1/{parent}/backups, so a rule can pin the agent to a single instance or backup and nothing else. You can allow read and provisioning calls while withholding destructive operations like revert and promoteReplica unless you explicitly grant them.