[Stack Monitoring] ESS 8.4+ deployments showing N/A for CPU throttling

[matschaffer]

If you run 8.4.3 (probably any 8.4) on ESS CPU throttling will gradually go to N/A. There are a few internal collection docs that provide data so you'll see 0 at first, but once metricbeat takes over you lose the insight (screenshot below shows N/A for two and 0 for one. A few intervals later they were all N/A.)

I can see some values in the data, so I'm thinking maybe this is a problem with the aliases in the index template.

[elasticmachine]

Pinging @elastic/infra-monitoring-ui (Team:Infra Monitoring UI)

[miltonhultgren]

It's a Metricbeat module bug, we don't copy over time_throttled_nanos from the API response to the event here.

So while the UI code tries to read the correct field here it will always be null.

The two fields you see are the counts which map to number_of_elapsed_periods and number_of_times_throttled.

[matschaffer]

Thanks! I saw those two coming in and though we had enough. Nice job catching the need for time_throttled_nanos as well.