Where the minutes go (30d)
Waste detected
Biggest wins first, each with the exact config fix.
Write-only cache (never restored) · CI
An `actions/cache` key includes the commit SHA with no `restore-keys`, so every run writes an entry no later run can restore. Key on your lockfile hash and add a `restore-keys` prefix.
- uses: actions/cache@v4
with:
path: <cache dir>
key: deps-${{ runner.os }}-${{ hashFiles('**/<lockfile>') }}
restore-keys: |
deps-${{ runner.os }}-Missing dependency cache · Build all targets from Linux (pinned soldr, zero host toolchain)
Add `Swatinem/rust-cache@v2` (the de-facto cargo cache), or an `actions/cache@v4` step keyed on `Cargo.lock`.
- uses: Swatinem/rust-cache@v2Workflow failing almost every run · Windows Docker Linux cross smoke
Fails or times out on nearly all recent runs, burning minutes to produce only red, and an always-red workflow usually gets ignored. Fix it, or disable the trigger until it's ready.
Missing dependency cache · Windows Docker Linux cross smoke
Add `Swatinem/rust-cache@v2` (the de-facto cargo cache), or an `actions/cache@v4` step keyed on `Cargo.lock`.
- uses: Swatinem/rust-cache@v2Missing dependency cache · thin-v2-verify
Add `Swatinem/rust-cache@v2` (the de-facto cargo cache), or an `actions/cache@v4` step keyed on `Cargo.lock`.
- uses: Swatinem/rust-cache@v2Missing dependency cache · Cook Size Gate
Add `Swatinem/rust-cache@v2` (the de-facto cargo cache), or an `actions/cache@v4` step keyed on `Cargo.lock`.
- uses: Swatinem/rust-cache@v2Also found: 7 more config fixes with negligible recoverable minutes: timeouts, retention, path filters
Premium runners (macOS / Windows) · CI
macOS bills ~10x and Windows ~2x a Linux minute. The cost estimate above assumes Linux, so your real spend is higher. Move any job that doesn't need them to `ubuntu-latest`.
jobs:
build:
runs-on: ubuntu-latest # ~10x cheaper than macos-latestPremium runners (macOS / Windows) · Setup Soldr Action
macOS bills ~10x and Windows ~2x a Linux minute. The cost estimate above assumes Linux, so your real spend is higher. Move any job that doesn't need them to `ubuntu-latest`.
jobs:
build:
runs-on: ubuntu-latest # ~10x cheaper than macos-latestCache without restore-keys · Perf Matrix
The cache restores only on an exact key match, so any lockfile change means a full cold download. Add a `restore-keys:` prefix line for partial restores.
key: deps-${{ runner.os }}-${{ hashFiles('**/<lockfile>') }}
restore-keys: |
deps-${{ runner.os }}-Scheduled workflow runs in forks · Perf Matrix
The schedule has no repository guard, so forks that enable Actions inherit the cron and burn their own minutes. Gate the job with `if: github.repository == 'owner/repo'`.
jobs:
nightly:
if: github.repository == 'owner/repo'
runs-on: ubuntu-latestFull-history clone (fetch-depth: 0) · Vendor state
checkout fetches the entire git history every run (`fetch-depth: 0`) and nothing in the workflow appears to read it. Remove the line; the default shallow clone is much faster on big repos. Keep it if a step genuinely needs history.
- uses: actions/checkout@v4
# fetch-depth: 0 removed — default shallow clone is enough herePremium runners (macOS / Windows) · Build all targets from Linux (pinned soldr, zero host toolchain)
macOS bills ~10x and Windows ~2x a Linux minute. The cost estimate above assumes Linux, so your real spend is higher. Move any job that doesn't need them to `ubuntu-latest`.
jobs:
build:
runs-on: ubuntu-latest # ~10x cheaper than macos-latestPremium runners (macOS / Windows) · Windows Docker Linux cross smoke
macOS bills ~10x and Windows ~2x a Linux minute. The cost estimate above assumes Linux, so your real spend is higher. Move any job that doesn't need them to `ubuntu-latest`.
jobs:
build:
runs-on: ubuntu-latest # ~10x cheaper than macos-latestCopy every fix as one block (6 snippets, annotated per workflow)
# Write-only cache (never restored) (applies to: CI)
- uses: actions/cache@v4
with:
path: <cache dir>
key: deps-${{ runner.os }}-${{ hashFiles('**/<lockfile>') }}
restore-keys: |
deps-${{ runner.os }}-
# Missing dependency cache (cargo) (applies to: Build all targets from Linux (pinned soldr, zero host toolchain), Windows Docker Linux cross smoke, thin-v2-verify, Cook Size Gate)
- uses: Swatinem/rust-cache@v2
# Premium runners (macOS / Windows) (applies to: CI, Setup Soldr Action, Build all targets from Linux (pinned soldr, zero host toolchain), Windows Docker Linux cross smoke)
jobs:
build:
runs-on: ubuntu-latest # ~10x cheaper than macos-latest
# Cache without restore-keys (applies to: Perf Matrix)
key: deps-${{ runner.os }}-${{ hashFiles('**/<lockfile>') }}
restore-keys: |
deps-${{ runner.os }}-
# Scheduled workflow runs in forks (applies to: Perf Matrix)
jobs:
nightly:
if: github.repository == 'owner/repo'
runs-on: ubuntu-latest
# Full-history clone (fetch-depth: 0) (applies to: Vendor state)
- uses: actions/checkout@v4
# fetch-depth: 0 removed — default shallow clone is enough hereEach snippet is representative. Merge into the named workflow files rather than pasting wholesale.
Live CI-health badge → 
[](https://gitspider.com/scan/zackees/soldr)Want this on every push?
This scorecard is a one-time snapshot. Install the free GitHub App to track this repo continuously: new regressions caught as they land, trends over time, on your public and private repos. Team adds the offending commit on the PR + Slack alerts.
Install & monitor this repo →💬 On Team, this same analysis posts automatically to every PR: the regression, and the exact commit that caused it, right where your team already looks. See plans →
Not ready to install? Get this report by email. No spam, unsubscribe anytime.