6 days ago
Service: cockpit digital ads (cockpit-ads-numeric-production)
Volume: “cockpit digital ads volume”, 500 MB, mounted on /app/data
The Railway dashboard shows a “High Volume Usage” alert indicating 85% of capacity used.
Connecting directly to the container, the following checks show actual usage well below that:
df -h /app/data → 5% used (18 MB / 434 MB), ext4 filesystem on /dev/zd12880
du -sh /app/data/* → 18 MB (only cockpit.db + cockpit.db-journal)
df -i /app/data → 1% of inodes used
No deleted-but-still-open file held by a process (lsof / /proc/*/fd)
No zfs/zpool tools available inside the container to check on the host side
Could you confirm whether a ZFS snapshot or host-side replication would explain this discrepancy, and release it if that’s the case?
1 Reply
6 days ago
This isn't caused by snapshots; snapshot space is never counted in the volume usage figure, which is tracked separately as Backup Usage. The figure in the dashboard is the space allocated to the volume on our storage layer, not the live data your filesystem reports, so it often sits well above df/du, especially for SQLite workloads that rewrite files often.
Freed blocks are returned to storage over time, but storage is allocated in 16 KB blocks while ext4 frees space in 4 KB blocks, so partially freed blocks can't be returned. The figure settles above what df shows rather than matching it. Your data is fine and there's nothing wrong inside the container.
We don't run a manual reclaim on a volume. Volume usage is billed on this allocated figure at $0.15/GB-month, so roughly 0.4 GB of allocation comes to about $0.06 per month.
Status changed to Awaiting User Response Railway • 6 days ago