Don't use systemd collector for node exporter
With enabled systemd collector for node_exporter prometheus receive metrics in more than the allowed timeout limit for this exporter.
As @bjk-gitlab suggested, we should not use systemd collector, but directly use up
metrics for these services and update corresponding alerts.
Services are:
Recording rules are:
related issue - https://gitlab.com/gitlab-com/infrastructure/issues/1408
/cc @bjk-gitlab
Edited by Ben Kochie