Node CPU alerts

This indicates there may not be enough CPU on the node in order to operate.

If there is an increase in CPU utilization on a single host, you may want to perform CPU profiling or look at profiles from our continuous profiling in Google Cloud Profiler.

For ad-hoc profiling, it is also possible to invoke the profiler on a process directly.

• For Go processes, you can profile via pprof.
• For Ruby processes, you can profile see: Ruby profiling.
• For C processes and kernel profiling, you can use perf, see also Redis - CPU profiling.