[Elastic Agent] Add feature flag to the system metricsets to control whether metrics failures mark agent as degraded #40543
Labels
Team:Elastic-Agent-Data-Plane
Label for the Agent Data Plane team
This is a follow up to #40542, where we made partial metrics collection keep the agent healthy to minimize customer disruption for non-fatal errors.
Once we are confident we have fixed the source of the most common partial metrics collection errors we should return to having agent report itself as degraded when metrics collection fails in our own tests as a quality gate. This allows new errors to be easily flagged to us before release.
The current known issues preventing us from doing this by default are:
We should add a feature flag to the system module or the system metrics that turns this functionality on or off, and turn it on in the Elastic Agent integration tests. Adding this as part of the system integration configuration allows it to be controlled through the configuration of the system/metrics input and system integration without needing changes to the Elastic Agent control protocol.
Acceptance Criteria:
The text was updated successfully, but these errors were encountered: