Built to scale: processing millions of records a minute

Screening one name is easy. Screening a bank's entire client book — hundreds of thousands of records — on a schedule is an engineering problem. Here's how Rahn Monitor handles scale.

~6,000 records/second per node

Our match engine sustains roughly 6,000 records per second on a single node, with median search latency under 200ms against the full 1.3M-record dataset.

Near-linear scaling with workers

Batch jobs are distributed across a pool of parallel workers. In production we run 32 workers, pushing throughput to roughly 11.5 million records per minute — enough to re-screen very large books well within a maintenance window.

Workers	Throughput
1	~0.36M / min
8	~2.9M / min
16	~5.8M / min
32	~11.5M / min

See the live scaling chart on our home page. Real-world rates vary with record complexity and match thresholds.