Grafana Protocol Detail dashboard confusing for nfs throughput

[mylesuyema]

graph labels such as:
read (read) read
read (read) write
lookup (namespace_read) write
lookup (namespace_read) read

[apecoraro]

What you are seeing is the name of the operation, followed by the class of the operation, followed by the word "write" if the data is going into the Isilon and the word "read" if the data is coming out of the Isilon.

You can change what that looks like by clicking on the title of the graph and selecting "Edit". Then on the editor screen in the Metrics tab you will see two metric queries, one which queries for the "mean(in_rate)" and one that is for the "mean(out_rate)". If you click the arrow button on the left side of the metric it will drop down and show the Alias by configuration which is currently set to "$tag_op_name ($tag_class_name) write" and "$tag_op_name ($tag_class_name) read" for the in_rate and out_rate respectively. So you can change that to something else to make it less confusing for you.

I'd be interested to know what you think it should be changed to?

[tenortim]

I think that "request" and "response" might be clearer labels here, since we're talking protocol operations from the perspective of the server. Does anyone have any additional thoughts?