#Has anyone undertaken any research into
tawdry zinc
Hello @azure veldt , we have an idea in the planning phase like this, but we haven't included it in the development roadmap yet. Here's the link for reference: https://permify.notion.site/Custom-Roles-7efcafec1b2a4a1b8e2c05a445b3be1a. I'd love to hear your thoughts!
azure veldt
"Tenants should, however, be able to customize the templates within boundaries set by the system's administrator to suit their needs." - In our case this probably fairly far off and we've been in business for ~10 years.
For segregation of data between tenants using the tenant column is good. But I am curious if performance could be increased by using native postgresql partitioning to make index sizes & table size when querying a particular tenant more performant etc (some our tenants have millions of resources). Also with larger tables split over N partitions you can parallelise vacuum & analyse operations as they are just their own tables etc.
Looking at the APIs, generally if I could create a schema that wasn't linked to any particular tenant and then create the relationship of tenant X to schema S, and say a new tenant is created I just create tenant Y and also link them to schema S; that would be sufficient for me/us I think.
Then when all data is stored for tenant X, their relationships are still stored with the association to their own tenant identifier then that seems reasonable.
dusky pier
hi @azure veldt, what is the difference between a global schema and same schema for all tenants? if we have a global schema what advantages will we have other than faster schema updates?
azure veldt
My understanding would be reduced memory usage across permify processes when the tenant count is in the thousands or tens of thousands.