Sure, swap is bad on steady state loads. But for transient loads it works
Unless you like playing "why did my service got killed this time" every once in a while
Sure, swap is bad on steady state loads. But for transient loads it works
Unless you like playing "why did my service got killed this time" every once in a while
There is no playing: the error is logged and quite clear, by the application or by the kernel.
Being consistent, predictable and as deterministic as possible is a key feature of reliable software, and paging into swap kills all that.
Paging into swap often has almost no perceptible effect in regular usage unless you are running software with microsecond guarantees on infrequently accessed processes (those that get swapped out).
Kubernetes as a platform recommends disabling swap completely, and you have to explicitly allow nodes to have a swap, otherwise it fails.
This is sane behaviour if you're dealing with a large cluster with a complex architecture you no single person could or should know all the ins and outs of. There is no "let's log on to the machine and see what's happening" when dealing with this types of architectures, even at smaller scale.
And a massive part of Go's target is exactly these modern architectures/workloads...
But still, you could end up with a case where your pods start using a little bit more memory (for a while) and getting killed by OOM