What should I use for a SaaS app that needs strong consistency?

If your SaaS app truly needs strong consistency, the safest default is:

Use a single strongly consistent database as your source of truth

Common choices:

• PostgreSQL — best general-purpose option for many SaaS apps
• MySQL/InnoDB — also solid
• SQL Server / Oracle — if you’re in those ecosystems

These give you:

• ACID transactions
• row-level locking
• constraints
• predictable reads/writes

If you need horizontal scale with strong consistency

Look at a distributed SQL database:

• Google Cloud Spanner
• CockroachDB
• YugabyteDB

These are built for:

• strong consistency across nodes/regions
• automatic replication
• higher availability than a single primary DB

Tradeoff: more complexity, higher latency, and sometimes more operational/cost overhead than Postgres.

What to avoid for strong consistency

Usually avoid using these as the primary system of record if consistency is critical:

• eventually consistent NoSQL systems
• caches as source of truth
• multi-writer architectures without a clear consistency model

Practical recommendation

For most SaaS apps:

1. Start with PostgreSQL
2. Use it as the system of record
3. Add read replicas, caching, and queues only for non-critical paths
4. Move to CockroachDB/Spanner only if you outgrow single-node/primary architecture or need multi-region writes with strong consistency

Rule of thumb

• Need simplicity and strong consistency? Postgres
• Need global scale with strong consistency? Spanner/CockroachDB/YugabyteDB
• Need fast writes and can tolerate staleness? NoSQL/eventual consistency systems

If you want, I can recommend the best option based on your expected traffic, region setup, and whether you need multi-region writes.