How to use Cloud Clone for datasets

Cloud Clone provides a more efficient way of duplicating datasets and is ideal for situations when:

• you want to run tests against real production data in a CI flow
• you regularly copy datasets from production for developing new features

Instead of exporting and importing a dataset with the CLI, you can have that process happen inside of Sanity's infrastructure which will be more efficient and reliable.

There are two methods of initiating and monitoring the cloning of datasets in the cloud: through the Sanity CLI or with the HTTP API.

Once a copy is successful, the new dataset will appear in Manage. Depending on the size of a dataset this may take hours, so we encourage monitoring the outcome of a copy using the jobId, discussed below.

The quickest way to begin developing with a freshly-copied dataset is to use the CLI.

By default, the CLI command runs the copy synchronously. If you don't want to wait for the process to be completed, you can use the --detach flag to skip the progress. It will log a job ID that you can use to watch the progress again with the --attach <jobId> flag.

Depending on the size of the dataset, skipping document history with the --skip-history flag can make the copy process significantly faster. In cases where document history is not important in the target dataset, this may be a flag worth considering.

# Syntax:
# sanity dataset copy
# sanity dataset copy <source-dataset>
# sanity dataset copy <source-dataset> <target-dataset>

# This command will ask for which dataset to copy and what to call the new dataset
sanity dataset copy

# This command will copy the production dataset and request a name for the new dataset
sanity dataset copy production

# This command will copy the production dataset into a new dataset named new-feature
sanity dataset copy production new-feature

# This command will initiate the copy between production and new-feature
# It will run in the background and not display progress while it works
sanity dataset copy production new-feature --detach

# This command will initiate the copy between production and new-feature
# It does not copy document history, speeding the copy action 
# at the expense of the history retention
sanity dataset copy production new-feature --skip-history

It's encouraged to use this feature instead of exporting/importing your data to another dataset. In most cases, this will be a faster method. On large datasets or datasets with a large number of assets and/or large assets, the process will take some time to complete.

If you'd prefer to use the HTTP API instead of the CLI, there are API endpoints for copying datasets and for monitoring copy completion.

PUT /v2021-06-07/projects/:projectId/datasets/:datasetName/copy

In order to start a copy, a PUT request is sent to the specific dataset's /copy endpoint.

https://api.sanity.io/v2021-06-07/projects/<project-id>/datasets/<dataset-name>/copy

The request needs to be authorized via a Bearer token, which can be generated from the Manage dashboard.

The body of the request must be an object containing the following fields:

1. targetDataset: Property to name the new dataset. The value must be consistent with dataset name requirements.

2. skipHistory: Boolean property which allows skipping document history while copying the dataset. It potentially reduces copying duration on datasets with large amount of edit history. Check the retention period to know how long a dataset's history is kept for.

{
    "targetDataset": "production-copy",
    "skipHistory": true
}

curl --location --request PUT 'https://api.sanity.io/v2021-06-07/projects/<project-id>/datasets/<dataset-name>/copy' \
  -H 'Authorization: Bearer <token-here>' \
  -H 'Content-Type: application/json' \
  --data-raw '{
    "targetDataset": "production-copy",
    "skipHistory": true
  }'

{
    "datasetName": "production",
    "message": "Starting copying dataset production to production-copy...",
    "aclMode": "public",
    "jobId": "jobIdString"
}

GET /v2021-06-07/jobs/:jobId

When you run a copy via the HTTP API, you'll receive a Job ID. This ID can be used to query the status of the clone job.

curl --location --request GET 'https://api.sanity.io/v2021-06-07/jobs/<jobid>' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer <token here>'

// Running
{
    "id": "jacsfsmnxp",
    "state": "running",
    "authors": [
        "authorId"
    ],
    "created_at": "2020-11-09T17:34:28.071123Z",
    "updated_at": "2020-11-09T17:34:28.144826Z"
}

// Completed
{
    "id": "jarrwsdptf",
    "state": "completed",
    "authors": [
        "authorId"
    ],
    "created_at": "2020-11-09T17:07:41.304227Z",
    "updated_at": "2020-11-09T17:08:30.457692Z"
}

GET /v2021-06-07/jobs/:jobId/listen

Each job has a /listen endpoint to allow you to monitor its status programmatically. Much like the static status endpoint, this endpoint accepts the Job ID that is returned by starting a copy action.

curl --location --request GET 'https://api.sanity.io/v2021-06-07/jobs/<jobid>/listen' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer <token here>'

While listening, event data will be sent back at intervals providing updates on the status of your copy. The response contains the event name as well as a JSON object containing information about the current status of the copy.

event: welcome
data: {"listener_id": "ladaicdbdo"}

event: job
data: {"job_id":"jacsfsmnxp","state":"running","progress":60}

event: job
data: {"job_id":"jacsfsmnxp","state":"running","progress":80}

event: job
data: {"job_id":"jacsfsmnxp","state":"completed","progress":100}