Someone mentioned to me yesterday that they want to run LED strips in their data warehouse that will tie trouble tickets with specific servers but couldn’t come up with a good idea. So last night I literally cobbled together a simple “healthy/failure” status to show them what it could look like on a single rack. Today, they already have additional ideas to be incorporated (as they should now that they see the possibilities.) I love my job and ability to incorporate my hobby with it. 
https://www.youtube.com/watch?v=GQs0tMFStfo
Nice! Also - damn, do servers go down that frequently?
No, I was manually sending FAIL signals to prove a point. 
What would be cool is if I can somehow display network traffic, real-time. Basically have one of the three LEDs flashing as packets go flying by to/from that server, the same way a NIC or switch does it. Or other hardware related pieces, such as disk activity, CPU usage … But I can’t do that without some software interface from each server to a central collector which then updates the string. Hrm … brain overflowing. Interfacing with SNMPd might be an idea, but it won’t be real-time.