Make sure that your investment in Database server hardware is sound, and represents an appropriate trade-off for your particular application. Only through rigorous benchmarking and testing can the current performance bottleneck be isolated.
2ndQuadrant can perform synthetic benchmarks representative of current or projected loads. Does a particular type of SSD represent a good choice, given the ratio of random I/O to sequential access seen for your workload? While we're on that topic, just how well regarded is your choice of drive by an inner cadre of database and hardware experts, particularly with regard to factors that cannot be directly measured, such as mean time between failures, or known issues with buggy though only intermittently problematic firmware?
Do you trust your hardware vendor to accurately represent the performance of their hard disks? 2ndQuadrant's considerable experience suggests that regardless of the vendor's reputation, you should not.