This Pint lint check still fails, even when I explicitly join on certain labels as in this PR. I think this might be a bug in Pint.

EDIT - fixed in https://github.com/cloudflare/pint/releases/tag/v0.71.0

I would like to be able to use without to cater for custom external labels that users might have configured on their metrics, so this seems like the lint check is a little too strict IMO.

Fixes #985

Rendered alert was tested manually on a fleet of clusters:

sum(increase(container_cpu_cfs_throttled_periods_total{container!="", job!="", }[5m])) without (id, metrics_path, name, image, endpoint, job, node)
  / on (cluster, namespace, pod, container, instance) group_left
sum(increase(container_cpu_cfs_periods_total{job!="", }[5m])) without (id, metrics_path, name, image, endpoint, job, node)
  > ( 25 / 100 )