Best practices for backing up and restoring datasets in Sanity.io

Sanity datasets and backup
• Whats the best way to backup a dataset in Sanity?
• Can I do it from the dashboard?
• How about automatic scheduled backups? Like say once a week?
• How do I restore from an backup?
• Can I backup a dataset called

production

and create a new datasett called

development

based on the

production

-backup?

1.

sanity dataset export

from the CLI or using the

/export

http api https://www.sanity.io/docs/export 2. I do not believe so
3. You could write a script that runs either of the above commands, but per my knowledge there is no automation other than what you build yourself
4.

sanity dataset import

is the easiest way. You can also use a client library, though that has some pitfalls: https://www.sanity.io/docs/importing-data 5. Yes. Run

sanity dataset export production

which by default will be saved to a file called

production.tar.gz

. Then you run and this will create a new dataset named

development

if one doesnt exist. If one does exist it will merge the two datasets. For more details look into the flags available on

import

with

sanity help dataset import

• Will the

sanity dataset export

command give me all data, including assets like images and files thats uploaded in the dataset?• What about the schema, is that a part of the export?
• Does that mean that I have to find some external cloud service to store my backups?

1. the export will include all user-generated documents, images and files, but will not include system documents that Sanity itself uses to function like custom access control ( https://www.sanity.io/docs/access-control#ce4a077926aa ) to my knowledge. 2. The schema isnt part of the dataset in the strictest sense. The schema is a concept that lives in the Studio and is how your Studio filters what documents are allowed to enter your dataset. Since it isnt part of the dataset it will therefore not be included in the export
a. Sidenote: Since only the Studio enforces your schema, you could also sidestep it by creating documents with the API or CLI that dont follow the schema. Be aware of this when creating documents this way
3. Yes, or some equivalent service according to your risk profile and data processor agreement

If I export

production

and then import it as

development

, do I then also have to do a graphql-deploy against my new

development

datasett for Gatsby and Sanity Studio to work agains the new

development

datasett?

I dont know, I dont use graphql in my projects. 🙂 I would guess so, but you should try and see what needs to happen. You could also see if there is some information here that is helpful: https://www.sanity.io/docs/cli-graphql

Thank you! I will try 🙂