问题
We are currently investigating, why our cluster isnt properly load balancing when a node is at 100% cpu usage.
Now when i look at the cluster via the service fabric explorer and click on Details, there are the metric information showing up. (e.g. Count, PrimaryCount, ReplicaCount etc).
But for some reason, the 2 default metrics "servicefabric:/_MemoryInMB" and "servicefabric:/_CpuCores" are not showing up there.
We aren't registering any custom metrics or anything on deployment/on runtime, which could lead to the metrics go missing.
Is there any explanation, why they are missing and is there a way to set them after cluster deployment? For example via Powershell?
We are running the cluster version: "6.0.211.9494"
Thanks in advance
回答1:
Okay, i fixed it by upgrading to the newest cluster version "6.1.467.9494"
The cluster was deployed with a json template quite a while ago, which registered the Cpu/memory metrics with lowercase letters in the beginning. The older cluster version only tried to read the metrics with upper casing.
The newest version fixed that and tries to read it ignoring the casing in the beginning.
来源：https://stackoverflow.com/questions/49237646/servicefabric-missing-default-metrics-servicefabric-memoryinmb-and-service