Observability alerts should expose a "reason"

[cyrille-leclerc]

Describe the feature:

Observability alerts should expose a "reason" that is a human readable description of a firing alert.
Reasons should describe

• The type of rule
• If possible on which resource/instance the violation of the rule have been observed
• Optionally an explanation of the threshold that have been configured
• Optionally an indication of the intensity of the violation of the rule

Example of "reason" for Observability alert

Rule Type	Reason
Error Count Threshold	Error count is greater than 100 (current value is 135) on shippingService
Transaction Duration Threshold	Latency is greater than 1500ms (current value is 1700ms) on frontend
Inventory	CPU is greater than a threshold of 75% (current value is 83%) on gke-eden-3-prod-pool-2-395ef018-06xg
Log Threshold	Log count with 'Log.level: error' and 'service.name: frontend' is greater than 75 (current value 122)

@XavierM suggests to look at the Security Row-render, it seems to be at the same problem.
See #91446
https://drive.google.com/file/d/1EcizAFktM91FoXICeyznFsVoYCRBhISz/view?usp=sharing

[cyrille-leclerc]

Issue moved to elastic/observability-dev #1405 via ZenHub