Data Structure / Schema Choice
Design the data model for a multi-tenant B2B SaaS task-management application on PostgreSQL 15, tuned for heavy read traffic
No card required
What this brief specifies
What's in scope
Eight distinct entity types have to share one schema: tasks, projects, teams, users, comments, attachments, custom fields, and an activity log.
Constraints
Multi-tenancy has to be single-database, not sharded, with strong isolation between tenants and an audit log required for every task change.
Scale to plan for
Up to 10,000 tenants, as many as 1 million tasks within a single tenant, and concurrent editing.
What makes this hard
Eight different entity types have to share one schema across up to 10,000 tenants without a hard per-tenant database boundary, since multi-tenancy here is single-database rather than sharded, meaning the same isolation guarantee has to be enforced consistently across every one of those eight tables, not just the obvious ones like tasks. The audit log requirement adds a second, different obligation, it applies specifically to task changes, so that one entity among the eight needs to preserve a history of state over time, not just its current row, while the rest carry no such burden. Layer the scale on top, up to a million tasks inside a single tenant with people editing concurrently, and a schema shape fine for a typical smaller tenant's dashboard can fail once one tenant alone reaches that row count while sharing the same database and indexes as thousands of others doing the same thing at once.
Related use cases
See how Flaibo works for Founder.
Frequently asked questions
The next multi-tenant schema that has to hold this many entity types under one isolation boundary, while surviving a tenant that scales far past the typical one, benefits from a data model stress-tested against its own worst-case tenant from the start, rather than one designed for the average case and patched once a single large tenant starts to strain it.