I remember a practice question where we had to analyze similar metrics, and I think a good indicator is the CPU load average being close to the number of cores.
I've got a strategy for this. I'll focus on finding the Ganglia metric that shows the overall resource usage across the 3 executor nodes. That should give me the best insight into how the VMs are being utilized.
I'm a bit unsure about this one. Ganglia metrics can be complex, but I think I remember something about CPU and memory utilization being good indicators. I'll have to double-check that.
Okay, let's see. The key is to identify which metric would indicate if the executor nodes are using the VM resources effectively. I'll have to review my notes on Ganglia.
Glenn
5 days agoDarrin
10 days agoLawrence
15 days agoRutha
20 days agoJosefa
26 days agoRuthann
1 month agoLaurel
1 month ago