Decreasing the logging interval is a possibility, but we’d rather not — 1GB of results per host times 5,000 hosts being consumed by multiple downstream services is a good bit more data than we’d prefer to throw towards these systems without stress testing.