I'm working on a botting software for an online game, and I'm very close to launch (4-6 weeks).
However, I can only run maybe one instance per two CPU cores. I don't want to get eager and shell out for some dedicated hardware just yet, so I was thinking about using Azure, Amazon, Google, or maybe some other cloud VPS to run my bots.
I'm trying to figure out how it works, though. I'm a complete newbie when it comes to cloud hosting. I've only used shared hosting plans on small web hosts...
1) Is the 'per hour' cost presented by Amazon and Microsoft a physical hour? Example: I run 5 bots (each using their own 'instance') over 4 hours using the Azure A2 plan, so I'd be charged $1.58 for 20 hours? Similarly, using the Amazon t2.medium plan, I'd be charged $0.94? As far as I can tell, that's the case with Amazon and Azure. (Haven't looked into Google at all.)
2) Is there any sort of CPU throttling or hidden cost for CPU utilization on lower-end plans (A plans on Azure or t2 plans on Amazon)? My bot software would use upwards of 150%-200% CPU usage (so almost full CPU utilization across both cores). So like the example in 1, I'd be charged $1.58 for 20 hours of 150%-200% CPU usage? Edit: Amazon seems to use some CPU credit system, but I can't find anything about Azure doing the same.
3) Is there any easy way to demo various plans without spending too much? Is it as simple as spinning up an instance and checking performance?
4) I've heard some horror stories about insane bills from hundreds of VMs being spun up, usually out of error or stolen tokens/keys/whatever-they-may-be-called. Assuming I only run a few VMs at once, is there anyway I can rack up an insane bill?
5) Any recommendations for a particular cloud VPS provider? Looking for personal experiences, if possible!
I'd be happy to provide any clarifications!