Hey! I've been using func build regularly across a few projects, and noticed something that might be worth a small feature wanted to check with you before writing any code.
The problem:
func build pulls builder images, run images, and creates buildpack cache layers on every run. Caching is clearly already happening (the CLI notes first builds are slower than later ones), but there's currently no way to see:
- how much disk space these images/layers are using
- which builder/run images are still active vs. left over from old projects
- whether a slow build is a cache miss or something else
Across multiple functions and projects, this accumulates with no built-in way to inspect or clean it up.
func builders prune reports disk usage per builder image, run image, and cache layer, with an option to remove ones no longer referenced by any local func.yaml.--cache-report flag on func build prints a per-layer cache hit/miss summary.
I checked open and closed issues/PRs for related prune/cache/disk-space work and didn't find anything in progress. If this was discussed or attempted before, I'd appreciate the context. Otherwise, happy to write a short design doc scoping this into reviewable pieces reporting first, prune as a follow-up.
Hey! I've been using func build regularly across a few projects, and noticed something that might be worth a small feature wanted to check with you before writing any code.
The problem:
func buildpulls builder images, run images, and creates buildpack cache layers on every run. Caching is clearly already happening (the CLI notes first builds are slower than later ones), but there's currently no way to see:Across multiple functions and projects, this accumulates with no built-in way to inspect or clean it up.
func builders prunereports disk usage per builder image, run image, and cache layer, with an option to remove ones no longer referenced by any localfunc.yaml.--cache-reportflag onfunc buildprints a per-layer cache hit/miss summary.I checked open and closed issues/PRs for related prune/cache/disk-space work and didn't find anything in progress. If this was discussed or attempted before, I'd appreciate the context. Otherwise, happy to write a short design doc scoping this into reviewable pieces reporting first, prune as a follow-up.