Store kubernetes clientset when waiting for apiserver
While looking at pprof data, I noticed that the line: ```go client, err = k.client(k8s) ``` takes around .9 seconds to complete. Previously, we were calling it on every iteration of the waitFor loop. Instead, if we store the clienset once we have it, we should be able to speed up this function. During local testing, this optimization reduced time spent waiting for the apiserver by 13 seconds on average.
Showing
想要评论请 注册 或 登录