(Node Stats) CPU Pressure

[stefnestor]

Description

👋 howdy, team!

Elasticsearch Node Stats currently emits os+process cpu instantaneous and load_* stats and cgroup's stats.

Will you kindly consider also reporting Metricbeat > System > Pressure fields for cgroup "cpu pressure" so that similar to heap vs memory pressure, we can get a better over time / recent average view of resource usage? E.g. fields like:

• system.process.cgroup.cpu.pressure.full.60.pct
• system.process.cgroup.cpu.pressure.some.60.pct

(Speculatively, this could be useful to add into Stack Monitoring or at least Elastic Cloud UI for users to be able to tell when their nodes are CPU strained. Seems maybe sub-ballpark of #88795.)

TIA for the consideration! 🙏

[elasticsearchmachine]

Pinging @elastic/es-data-management (Team:Data Management)

[consulthys]

this could be useful to add into Stack Monitoring

Definitely, subscribing to this issue so we can tackle that whenever that's available.