Foundgine Performance

CoffeeBeanery PostgreSQL graph benchmark — 12 August 2026.

This benchmark measures the Provider stage against PostgreSQL.

Three independent successful runs were performed against a deterministic PostgreSQL graph workload.

Workload

Customer
  → CustomerBankingRelationship
      → Contract
          → Transaction

Fixture:

• 1,000 customers
• 4,000 relationships
• 12,000 contracts
• 48,000 transactions
• concurrency 1, 8, 32
• 10-second measurement per case
• 3-second warm-up
• 5-second request timeout

Query result

At concurrency 32:

The two Foundgine configurations are shown independently so the effect of provider-plan caching can be evaluated without introducing an unrelated implementation comparison.

Reliability

The three successful runs reported:

• 0 application errors
• 0 request timeouts
• 0 cancelled requests

Mutation results

Mutation performance is more variable. The benchmark supports the conclusion that Foundgine can perform well at higher concurrency, but mutation performance should not currently be presented as the primary performance claim.

What this proves

The strongest evidence is for:

read/query execution over a relationship-heavy PostgreSQL graph workload.

The results consistently show substantially higher query throughput and lower p95 latency in this controlled workload.

What this does not prove

This is not a universal benchmark of every PostgreSQL schema, query shape, mutation workload, or hardware configuration. Results depend on the workload, schema, provider version, host, fixture, and implementation version.

The appropriate claim is:

Foundgine demonstrates strong query performance for this relationship-heavy graph workload, with provider-plan caching measured separately.

Full benchmark source data is kept in the repository alongside this page.