Measure the cost of various stages of issuing an audit log entry.

1. Construction of client metadata
2. Construction of impersonation metadata
3. Synthesis of BSONObj containing full audit record
4. Application of auditFilter
5. Synchronous write to disk of selected audit entry

Ideally create this as a benchmark framework to be reused across multiple benchmark runs with varying amounts of metadata, log entry size/complexity, BSON/JSON output, mongo/OCSF schema, nil/basic/complex filtering, etc...