Yeah I've been looking at the Supermicro for just that setup.
Supermicro is also my favorite hardware vendor (in those cases where I have the freedom to choose hardware).
Interesting you went with SSDs for DB. Speed, but the longevity is my concern there.
It depends. To begin with, SSDs can live for a very long time, if a lot of the workload is reads, and only a small fraction is writes. Such workloads do exist, in particular in database applications. Another thing that modern applications do is to optimize their write traffic to prevent "false write sharing" on SSDs, by doing interesting logging techniques.
And remember that hard disks don't live forever either. Matter-of-fact, modern hard disks are specified to have a maximum data traffic per year; for most models I've seen recently, that specification is 550 TB per year (so a 10TB disk can only be read 55 times per year, or roughly once a week). Note that for SSDs the endurance limit is determined by writes, for spinning rust by total traffic, which makes a huge difference.
And ultimately, if an application simply needs the write speed of SSDs, then endurance becomes secondary. There are customer environments where SSDs need to be replaced every few years. Obviously, this is expensive (service personnel needs to be scheduled, spare parts need to be stockpiled and shipped, support contracts need to be written and paid for), but having a broken computer may be more expensive. Obviously, this only makes sense with some sort of RAID mechanism, and in most cases only if the replacement can be done "hot", on a live system.
Think of it this way: If you build a high-end storage server, the computer itself (the CPU box, with PCIe disk controller and network cards) may cost roughly $10K, depending on how much memory you put in there. The external disk enclosures (say 5 extra rack-mount JBODs with room for 80 disks each) can easily add another $50K. The disks themselves start at $100K (if you use reasonably inexpensive enterprise-grade near line drives), all the way up to quarter million or half a million (if you use flash storage like SSDs). Add to that some software, vendor overhead, and support contracts, and you have a very expensive single system, with most of the cost in the storage disks themselves. But why do people buy these very expensive systems? Because they need them to make money. Usually that means that the data on the server is more valuable than the cost of the server, by a large margin. In the overall scheme of a typical corporation, the cost of IT infrastructure is still a minor factor, even though a high-end server cluster (like the one SirDice described above, or like the storage server I'm talking about) can cost as much as a private residence.
