GitHub Actions runner out of disk mid-job: "No space left on device" on a disposable machine
verbatim errorERROR: failed to run docker build: No space left on device
# standard runner: ~14GB free after setup, consumed by images + artifacts + build cache
##[error]The runner has received a shutdown signal...
Problem
A Docker-heavy workflow started failing mid-run after a base image update:
ERROR: failed to run docker build: No space left on device
# standard runner: ~14GB free after setup, consumed by images + artifacts + build cache
##[error]The runner has received a shutdown signal...Nothing in the workflow changed except a prebuilt image growing by 3GB. The ubuntu-latest runner's disk had become a scarce resource.
Root cause
The standard runner has roughly 14GB of genuinely free disk once the OS, toolchain and pre-installed images are counted. Chrome/Playwright installs, docker layers, dependency caches, and test artifacts all land on the same disk, and jobs that used to fit "with room to spare" no longer do. There is no persistent state to blame — the machine is disposable — so the fix is entirely about how much the job brings with it.
fix preview — first 2 of 3 lines (yaml), truncated:
- name: Disk report
if: failure()
… 1 more line in the fix🔒 the fix — including 3 code blocks — is members-only. $1/mo unlocks everything.