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
GCP Janitor — a new command for reviewing what a project is actually running and paying for.
Resource inventory across VM instances, persistent disks, snapshots, and static IP addresses, with name filtering and detailed per-instance views.
"Zombie" detection that surfaces likely waste: unattached disks, stopped VMs past a configurable idle threshold, orphaned snapshots, and unused static IPs.
Mark any flagged resource as an exception (with an optional reason) so it's excluded from future scans. The review is non-destructive — nothing is deleted for you.
Export the filtered VM inventory to a timestamped CSV file in the current directory.
Compute Engine support — lists instances, disks, snapshots, and addresses across the project, resolves machine types, and provides detailed per-instance information.
Configurable settings via a settings.json file (local directory or global config), including list size limits and the stopped-VM idle threshold used by the janitor (default 30 days).
Improvements
Unified, colorized CLI output with consistent info / success / warning / error messages and refreshed banner and update-notification rendering.
Resource status is now color-coded for quick scanning (running / in-use in green, stopped / terminated / unattached in red, everything else in yellow).