This change tries to prevent issue when the host is overloaded.
Let's go !
Pull-Request has been merged by tdecacqu
Could we have that memory value as a service parameter ? Could it be an idea to have the capability to limit the amount of containers that k1s get up and running at the same time ?
Yes, perhaps using a K1S_POD_MEMORY_LIMIT environment variable would be the easiest. I wary adding any dynamic logic to the service since it can be hard to predict what will be the memory usage, of lets say, 5 tox jobs started at the same time.
This change tries to prevent issue when the host is overloaded.