Artifacts and Caching

Overview

Configure artefact upload/download so a build job produces a shareable package a test job consumes, and add a lockfile-keyed dependency cache without treating cache as a correctness guarantee.

Artefacts (GitHub spelling in the product UI: artifacts) are job outputs GitHub stores for download, retention, and sharing across jobs in a workflow run — packages, binaries, test reports, and logs. Caching restores dependency directories between runs to cut install time; cache is best-effort and must never replace pinned lockfiles. Confusing the two causes flaky pipelines and bloated storage bills.

This is a core tutorial in Module 6 · Artifacts & Caching of the REBASH Academy GitHub Actions for Cloud & DevOps Engineers series — written for Cloud, DevOps, Platform, and SRE engineers.

Prerequisites

• Secrets, Variables, and OIDC

Learning Objectives

By the end of this tutorial, you will be able to:

• Upload artefacts with actions/upload-artifact and retention
• Download artefacts in a later job with actions/download-artifact
• Configure actions/cache (or setup-* cache) with a lockfile key
• Distinguish artefact vs cache vs container registry image
• List pitfalls that cause stale or missing caches

Architecture

This topic’s control points and relationships are shown below.

Theory

What it is

GitHub Actions separates what you ship between jobs from what you hope to reuse for speed:

Mechanism	Purpose	Guaranteed?
Artefacts	Persist outputs for later jobs / humans	Yes (within retention)
Cache	Speed dependency or build dirs across runs	No — miss, evict, or empty
Package / container registry	Immutable release distribution	Via your registry policy

Upload artefact actions package paths from the runner and store them against the workflow run. Downstream jobs download by name (and optional pattern). Retention defaults apply; set shorter retention for intermediate builds and keep release artefacts on tags or push immutable images to GHCR / another registry.

Cache stores a keyed blob (for example ~/.npm or ~/.cache/pip). Keys usually hash package-lock.json, poetry.lock, or go.sum so installs invalidate when dependencies change. Restore keys allow partial matches (same OS, older lockfile hash) as a softer fallback. Official setup actions (actions/setup-node, setup-python, …) often wrap caching for you.

Why it matters

Slow pipelines waste runner minutes and delay review feedback. Wrong cache design causes “works on my branch” builds when a stale cache hides a missing lockfile or native binary mismatch. Artefacts make builds auditable: the same tarball that passed tests can be promoted or attached to a release. Platform teams set retention and size policies so Actions storage does not become an unbounded object store. In regulated environments, knowing which artefact SHA was deployed matters as much as the Git commit.

How it works

1. A build job writes files under a path (for example dist/) and uploads them with a stable artefact name.
2. A test job declares needs: [build], downloads that artefact, and runs checks against the exact bits.
3. An optional cache step restores a dependency directory using a key derived from the lockfile; the job still runs install if the cache is cold.
4. Failed jobs can still upload (if-no-files-found, if: always()) when you need logs or reports.
5. Expired artefacts disappear; caches may be evicted — never assume either is permanent storage.

Prefer short retention for intermediate binaries; publish releases to a registry or GitHub Releases rather than relying on 90-day artefact downloads.

Key concepts and comparisons

Concern	Artefacts	Cache
Correctness	Part of the pipeline contract	Optimisation only
Scope	Same workflow run (primarily)	Cross-run within repo/branch scope rules
UI	Downloadable from the run	Opaque speed-up
Security	Can contain secrets — scope carefully	Same if you cache credentials

Strategy	Key idea
Exact lockfile key	Invalidate on dependency change
Restore-keys prefix	Warm partial hit when exact misses
Per-OS in key	Avoid Linux cache on macOS runners
Segment by job	Prevent test caches poisoning build caches

Common pitfalls

• Treating cache as a substitute for committing lockfiles.
• Caching the entire workspace including .env or cloud credentials.
• Omitting retention and filling storage with huge node_modules artefacts (artefacts are the wrong tool for that — use cache or registry).
• Expecting downloads without needs: ordering or mismatched artefact names.
• Sharing mutable caches across privileged and untrusted jobs without isolation.
• Confusing job artefacts with images in GitHub Container Registry (GHCR).

Hands-on Lab

Create a workspace for this tutorial.

mkdir -p ~/rebash-github-actions/module-06/.github/workflows && cd ~/rebash-github-actions/module-06/.github/workflows

Focus: actions/cache patterns and upload/download-artifact across jobs

Step 1 – Cache + artifact workflow

mkdir -p .github/workflows src
echo 'print(1)' > src/app.py
cat > .github/workflows/artifacts.yml << 'EOF'
name: Artifacts and caching
on: [push, workflow_dispatch]
permissions:
  contents: read
jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-python@v5
        with:
          python-version: "3.12"
          cache: pip
      - run: |
          pip install pyyaml
          mkdir -p dist && cp src/app.py dist/
          echo "revision" > dist/REVISION
      - uses: actions/upload-artifact@v4
        with:
          name: dist
          path: dist/
  verify:
    needs: build
    runs-on: ubuntu-latest
    steps:
      - uses: actions/download-artifact@v4
        with:
          name: dist
          path: dist
      - run: test -f dist/REVISION && cat dist/REVISION
EOF

Step 2 – Confirm cache and artifact actions

grep -E 'upload-artifact|download-artifact|cache:' .github/workflows/artifacts.yml

Final step – Cleanup note

# Keep ~/rebash-github-actions/ for later tutorials

Validation

• Lab commands run under ~/rebash-github-actions/module-06/.github/workflows/
• You can explain each Theory section in your own words
• You used modern tooling where it applies to this topic
• You can describe one production failure mode for this topic

Code Walkthrough

Production practice for Artifacts and Caching always combines:

1. Inspect before you change (status, plan, logs, dry-run)
2. Prefer reversible, documented changes (Git, IaC, drop-ins, version pins)
3. Capture evidence (command output, pipeline logs) for handovers
4. Prefer current tools and APIs over legacy shortcuts
5. Least privilege — escalate credentials only when required

Keep runbooks short enough to follow under pressure. Automate checks; keep humans for judgement.

Security Considerations

• Treat credentials and tokens for github-actions as privileged — never commit them
• Prefer short-lived auth (OIDC, roles, SSO) over long-lived keys
• Validate blast radius before apply/deploy/delete operations
• Restrict who can approve production changes
• Collect audit logs; limit who can read sensitive traces

Common Mistakes

Treating cache as a substitute for committing lockfiles.

Validate assumptions against the Theory section and official docs before changing production.

Caching the entire workspace including .env or cloud credentials.

Lab shortcuts (open security groups, admin roles, skip approvals) must not ship unchanged.

Changing production without a rollback path

Always know how to revert (previous artefact, prior release, state rollback, DNS failback).

Best Practices

• Encode Artifacts and Caching changes as code and review them in pull requests
• Pin versions (images, modules, actions, provider plugins)
• Separate environments with clear promotion gates
• Alert on symptoms with runbooks attached
• Destroy lab resources; tag everything with owner and expiry where possible

Troubleshooting

Symptom	Likely cause	Fix
Auth / permission denied	Wrong identity, policy, or scope	Check caller identity, roles, and least-privilege policies
Timeout / no route	Network, DNS, security group, or endpoint	Trace path, DNS, and allow-lists before retrying
Drift / unexpected plan	Manual change or wrong state/workspace	Reconcile desired vs actual; avoid click-ops on managed resources
Pipeline/job red	Flaky step, cache, or missing secret	Read failing step logs; bisect recent workflow/config changes
Cost spike	Idle load balancer, NAT, oversized compute	Inventory billable resources; stop/delete labs promptly

Summary

Artifacts and Caching is essential for Cloud and DevOps engineers working with github-actions. Practise the lab until the inspection and change path is muscle memory, then continue the track.

Interview Questions

1. Cache vs artifact — which is authoritative for build outputs?
2. Cache restores but builds still slow — what else matters?
3. How do you pass files from job A to job B reliably?
4. What security caution applies to caches?
5. When should artifact retention be short?

Sample answer — question 2

Confirm upload/download names match and the producer job succeeded. Caches are best-effort acceleration, not a contract between jobs.

Sample answer — question 4

Do not cache secrets or writable shared directories across untrusted branches without careful keying.

Related Tutorials

• Course overview
• Docker Pipelines with GitHub Actions

References

• Storing workflow data as artifacts
• Caching dependencies to speed up workflows
• upload-artifact · cache