Has anyone noticed that oracle keeps changing the idle requirements for compute instances?
I just checked their docs:
- CPU utilization for the 95th percentile is less than 20%
- Network utilization is less than 20%
- Memory utilization is less than 20% (applies to A1 shapes only)
In May it was 15% and in March it was 10%.
Is there anyway I can keep my compute instance without having them keep reclaiming it?
How do they actually get that information (particularly memory utilization)? Do they rely on their agent that’s pre-installed (but can be uninstalled)? At least in their web interface it doesn’t show any of that utilization for my instances (one is Ubuntu with their agent uninstalled and the other one is NetBSD).
I don’t know, wouldn’t the Hypervisor be able to track resources usage by itself without anything else?
At least for memory usage the hypervisor wouldn’t be able to tell the difference between memory merely used as cache vs. memory actually used by the software running on the machine (and OSes will usually just use any otherwise unused memory as cache, so you will likely see some inflated memory usage)
I don’t think they care about the fine details. The just drag a slider and it tells them it would kick out this many free instances, and someone says “ok let’s go with that”.