No benchmark in MongoDB.Driver.Benchmarks enables logging, so the log truncation path is entirely unmeasured and the fix in CSHARP-6232 has nothing guarding it against regression. DRIVERS-2804 calls this out directly: the problem will not be caught in testing without specifically benchmarking large documents with logging enabled, which is why the C# perf boards never surfaced it.

The benchmark needs to cover documents large enough that a full serialize-then-truncate is visibly more expensive than truncating early, with the logger enabled at Debug for the command component.

Scope is a driver-local benchmark, not DriverBench. Extending the standardized benchmarking spec was considered and rejected: a logging-enabled score is dominated by the language's logging framework and sink rather than by the driver, so it cannot serve that spec's cross-driver comparison purpose, and adding tasks perturbs the composite score weights.

Follows from CSHARP-4932 and guards CSHARP-6232.