rationale: in its current state, it adds a lot of stress yet noone seem to look into its results. it also requires package rebuild and redistribution (https://gitlab.com/gitlab-com/infrastructure/issues/3471), so I propose we tear it down from gprd environment, and iterate on including it later according to production readiness requirements, with the aspects like system load and log rotation addressed _before_ its installed everywhere. /cc @gl-infra