Can I use one convex backend for let's say 1.000 customers * 10M rows
#How scalable is convex for multi tenant databases with millions of rows?
winged summit
daring compassBOT
Hi @winged summit. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
Convex is explicitly designed to scale to very large datasets and high traffic, but the knowledge sources only give partial, not exact, answers for your specific “1,000 tenants × 10M rows” scenario.
What we do know from the docs
- Paid plans “have no hard resource limits – they can scale to billions of documents and TBs of storage.” [State limits]
- The platform is built to handle large-scale workloads; the team has “built exabyte-scale storage systems before” and designed Convex “with scale in mind.” [Horizontally scaled]
- Function execution has been horizontally scaled via the Funrun service, allowing Pro customers to run 10× as many functions concurrently, and Funrun itself can be scaled out as needed. [Funrun impact]
- Professional plans include 50 GiB storage and can go beyond that with per‑GiB pricing; there are no documented hard caps on total document count beyond storage. [State limits]
- Convex emphasizes that you should expect to scale to “thousands to millions of users” and that scaling techniques (indexes, pagination, data segmentation) are the normal path as your app grows. [Queries that scale; Operational maturity]
So from the official limits and positioning, a single Convex deployment is intended to handle very large, multi-tenant applications with many millions (and even billions) of documents, assuming:
- You stay within storage/bandwidth/function-call limits for your plan.
- Your queries are written to scale (proper indexing, pagination, etc.). [Queries that scale]
What’s not explicitly documented
- There is no explicit example or guarantee in the sources that “1,000 tenants × 10M rows each” (i.e. ~10B rows) in a single deployment is supported or recommended.
- There is no specific multi-tenant sizing guideline (e.g. “X tenants × Y rows per tenant per deployment”).
Given that the docs say “billions of documents” and “TBs of storage” are supported on paid plans, your order of magnitude (10M rows per tenant, 1,000 tenants) is at the very high end of what’s implied, and whether it’s practical will depend on:
- Actual document size (how many TB that becomes).
- Access patterns and query design (indexes, sharding patterns, etc.).
- Required latency and concurrency.
Practical takeaway
- Yes, Convex is designed to be used as a single multi-tenant backend and can scale to very large numbers of rows and users. [State limits; 70‑module backend]
- The knowledge sources do not explicitly confirm that a single deployment is tested/supported at the exact scale of 1,000 tenants × 10M rows each.
- For that level, you should treat it as an “enterprise-scale” scenario and talk directly with the Convex team (via support or Discord) to validate capacity planning and architecture for your specific workload. [State limits; Queries summary]