[Alex R (alexr17)]

API pods need an explicit opt-in to collect container CPU metrics in compatible API images.
Set ENABLE_CONTAINER_CPU_METRICS=true in the shared API ConfigMap, with opt-out through api.extraEnvVars and no image or chart version changes.

Input	Intended behavior
AWS, Google, Azure; CI and example values	Enable the shared API setting across providers.
Workload isolation off/on	The default API deployment and the default, ingest, and background pools inherit the setting.
Enabled → disabled → enabled	An explicit api.extraEnvVars value of "false" disables collection; remove it or set "true" to enable again.
Reused values; missing new maps; partial overrides	No new values or maps are read, so retained values receive the setting and existing environment overrides still win.
Missing, null, empty, or zero feature values	There is no new feature value to interpret; the existing environment override accepts string values.
Release tags with/without v, full/short/all-digit SHAs, prereleases, custom tags	No new image parsing or version gate; existing startup-gate requirements still apply. Older API images ignore the setting.