Gao PeiLiang
Gao PeiLiang
+1 on this , I also found it on etcd cluster master node , when add etcd3_alert.rules ..  it will cycle five mintue ... but we can't find something...
/remove-lifecycle stale
when I special one node , QL is same as `sum((avg(irate(node_cpu_seconds_total{instance=~"$node", mode="system"}[5m])) * 100))` I think is the $node argument is not correct ......
https://github.com/grafana/kubernetes-app/issues/13
> > I have see 0.3 release an network qos plugin ,, is there any doc about it ? > > we just see set cgroup net class id,, but...
> Right now, only Active Pods are added to the Endpoint, and Pods won’t be included when `publishNotReadyAddresses` is set to false. Yes, It like source code said ! .......
I think should focuse on it has different action of filter "k8s service packet" when use iptable mode with ipvs mode , when turn iptables to ipvs someday,, it will...
 pod -> IPVS VIP service -> pod ,,, netfilter hit 1 -> 2. (it is different with pod -> iptables VIP about netfiler path) the (IPVS VIP) last action...
> 当前版本网络信息会缓存，故还不支持两个webhook server。今年应该会进行改造 我也发现这个问题了， 好像默认安装 副本 就是 2 个，所以总有很多人遇到这个问题，，