An EC2 database uses a 500-GiB gp3 EBS volume. Measurements show sustained throughput demand near 200 MiB/s, below the instance EBS bandwidth limit. The volume has baseline gp3 settings, and storage capacity and IOPS are sufficient. Which change most directly addresses the throughput bottleneck?

1. Move the volume to st1 without considering the database access pattern.
2. Increase the gp3 provisioned throughput to at least the measured requirement without increasing storage capacity unnecessarily.
3. Increase the instance size even though the measured instance EBS bandwidth is not exhausted.
4. Increase only the gp3 volume size because throughput always scales automatically with capacity.