Local SSD is what makes CI feel fast
Disk types for GCP runners
Which GCP disk types to use for runner scratch space, and why local SSD matters more than machine size for most jobs.
A surprising share of CI time is disk-bound rather than CPU-bound: npm writing tens of thousands of small files, Cargo unpacking crates, Docker extracting layers. Network-attached disk makes all of it slower in a way no CPU graph shows.
| Disk type | Suitable for | Notes |
|---|---|---|
| Local SSD (NVMe) | Runner scratch, build directories | Fastest; ephemeral by design, which suits CI |
| pd-balanced | Boot disk | Fine for the OS, not for build output |
| pd-ssd | Fallback where local SSD is unavailable | Slower than local, and billed for provisioned capacity |
| pd-standard | Nothing in CI | Latency makes file-heavy steps painful |
Your next build could be twice as fast, at half the price
Start free. Migrating away is the same one line, and we publish that diff too.