Using a "Non-zero min_interval" causes warning-log per target

[KumaNeeko]

Just upgraded from 0.16.0 to 0.18.2, and noticed a bunch of warnings on startup:

time=2025-09-15T13:11:16.625Z level=WARN source=collector.go:67 msg="Non-zero min_interval, using cached collector." logContext="job=<job>,target=<target>,collector=<collector>" min_interval=10m

I have somewhere around 60+ targets, so this caused quite the flood of WARN-logs, but from what I can tell, they are harmless?
We have specifically configured a min_interval, to reduce the amount of times the database is actually queried.

It seems to have been updated in the following commit from INFO to WARN:
fd44637#diff-f1c80b360812fd52bef43d56f3c4ef40853f02711810857332ccedfd9f463f79R67

I guess my question is, if I've missed something in the configuration for this value, that would constitute this being a WARNing?

[burningalchemist]

Hey @KumaNeeko, Yeah, the warnings are just for the awareness, to signify that the metrics served are cached, so it might not reflect the actual state of the backend.

There are no other side effects, the adjustment was mainly on the log level of this particular message.