We'd like to reduce the frequency of Pod replacement. Currently, we replace Pods every time VPA suggests different values.
We can reduce, for example, by:
When VPA suggests scaling up, we scale up ASAP (otherwise might dangerous). But, when VPA suggests scaling down, we don't scale down very often. If VPA strongly wants to scale down, it should keep suggesting, and we can reduce the value in such case only.
We'd like to reduce the frequency of Pod replacement. Currently, we replace Pods every time VPA suggests different values.
We can reduce, for example, by: When VPA suggests scaling up, we scale up ASAP (otherwise might dangerous). But, when VPA suggests scaling down, we don't scale down very often. If VPA strongly wants to scale down, it should keep suggesting, and we can reduce the value in such case only.