Right after a deploy this afternoon, every host in a service's pool starts running low on disk, all at once,…

Question

Right after a deploy this afternoon, every host in a service's pool starts running low on disk, all at once, and disk usage is climbing fast — not over days, but over minutes. The application is still serving traffic but is slowing down. You notice the new release turned on debug-level logging. How do you triage and stop the disk from filling?

Prep Room · Answer

Two clues immediately separate this from a slow capacity drift: it hit every host in the pool at once, and disk is climbing over minutes, not days — and it…

Disk fills after deploy