sigh they should be using elastic solutions to this. If they ran on a kubernetes cluster, they could build a service that spins up a container for every video that comes in.
Sounds like they failed to factor in scalability.
As someone who works on software tools that has scaled from a single 8 core instance to a cluster of 200 64 core boxes, I’ve got a bit of experience in the area. Solutions that tend to fail are the ones that aren’t able to handle the scaling. It’s not so much “did we design a good product?” That’s easy. “did we design a product that won’t choke when it gets a lot of attention” is where the real money is.
The problem is that their hiring process is a bunch of bullshit. Fill out a google form with your experience, then 6 months later you get an email asking for resumes. You’re not going to get anyone good that way. You really need to use or hire a recruitment officer.
This is the wrong mentality but the correct problem.
It needs to be built so it can scale 1000x in 30 minutes.
If you build a solution with proper ISC (see here: http://howtocookmicroservices.com/communication/), you can get insane scalability at the drop of a hat. People need to stop thinking about platforms and start thinking about tasks. You spawn workers for each task. A server doesn’t provide specific services, but a platform to run generic services. One day server 5 may be hosting a transcode engine when the next day it might be hosting an nginx proxy or auth service.