E-commerce · Incident response
A full disk took an online store down. It was back the same afternoon
An online store stopped working because its server's disk was full. We freed 340 GB, disk use fell from 89% to 5%, and the store was selling again the same afternoon.
The situation
An online store went down. The cause was simple and common: the server’s disk was full. When there is no space left, the database cannot write, sessions cannot be saved and orders cannot be placed.
Full disks rarely happen overnight. Something grows quietly for months until one day there is no room left.
What we did
Find what is taking the space. Before deleting anything, we measured where the space had gone, so we removed what was safe to remove and nothing else.
Free space carefully. In total we freed 340 GB. Disk use went from 89% to 5%. Following our usual rule, removed files go to quarantine first, not straight to deletion, so nothing important is lost by mistake.
Bring the store back. With space available again, the store came back up the same afternoon.
Outcome
The store was selling again within hours, and the server went from nearly full to almost empty.
A full disk is one of the easiest outages to prevent. Disk use alerts, regular clean-up of whatever grows, and backups that go to separate storage instead of the same disk keep it from happening again. That is part of what we set up in managed operations.