Run_loop_busy metric values

We are started to dig deeper to find a reason why our clients are slow and we started to take a look on status json and found a metric run_loop_busy.

On storage nodes we have it almost 0.3. Does this mean that storage servers are freezing for almost 0.3 seconds to process something? Is this normal or we should scale something?

Run loop busyness is a measure of what fraction of the time the run loop thread is doing work. A value of 0.3 means it’s only active 30% of the time, and so it should have a good bit of extra capacity.

1 Like