Support better multi-tenancy with @tenant transformer

Is your feature request related to a problem? Please describe. Multi-tenancy (MT) implementation support using Amplify is very limited. The only MT strategy that is possible is to have a tenantId column in tables. With 100+ tenants, table sizes might explode. Making lookups slow and deleting a tenant data complicated.

A better alternative is to have tenant specific tables using prefixing: tenantid1_posts, tenantid_2_posts … . This can possibly be achieved with a @tenant transformer with something like:

type Blog @model {
  id: ID!
  name: String!
  posts: [Post] @connection(name: "BlogPosts")
}
type Post @model @tenant {
  id: ID!
  title: String!
  blog: Blog @connection(name: "BlogPosts")
  tenantid: String
}

such as when a tenantid is provided, the entry will be created in a separate tenantid prefixed table/collection.

About MT strategies https://d0.awsstatic.com/whitepapers/Multi_Tenant_SaaS_Storage_Strategies.pdf

[undefobj]

@mdegrees We are looking at more multi-tenancy support this year, however rather than look at a specific implementation could you uplevel a bit and go into detail around your use cases and scenarios? This would allow us to look at if there is a point feature causing problems here or a larger design that is cross cutting categories which needs implementation. Thanks.

[undefobj]

No I don’t have an example of updating records handy, but you should be able to do this using Lambda and the DynamoDB document client.