You're welcome. We've been running some benchmarks, and there were some optimizations being done for multi-tenancy, using a thread-local pooled optimizer. But for the pooled optimizer, if you increase the increment size, it takes around 150 microseconds to call the database sequence which is really fast.