You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
custom metrics: applications.custom_stale_after to tune how long a group
keeps reporting after its producer stops (default: a few collection intervals)
Fixed
storage: tolerate unclean-shutdown holes in tier reads
custom metrics: stop displaying a group's last values once its producer stops
pushing — stale feeds now leave a gap instead of a frozen line
custom metrics: dedup repeated metric names per message (last value wins),
preventing duplicate Prometheus series and rejected scrapes
custom metrics: data race on the debug flag; surface oversized/failed socket
reads instead of dropping them silently
Changed
custom metrics: reap idle socket connections after 5m; emit values in config order