feat: Add continuous deployment pipeline with manual version control

[MoonBoi9001]

Summary

Implements a comprehensive continuous deployment (CD) system with manual version control, enabling controlled releases with semantic versioning and automated Docker image publishing to GitHub Container Registry.

Motivation and Context

Previously, Docker image builds and releases were manual processes without version tracking. This CD pipeline provides:

• Controlled releases: Manual trigger prevents accidental deployments
• Semantic versioning: Proper v0.0.1, v0.0.2, v0.1.0 version tracking
• Multi-architecture support: Builds for both amd64 and arm64
• Production readiness: GitHub Container Registry publishing with proper tagging
• Release documentation: Automated changelog generation and GitHub releases

Changes Made

✨ feat: GitHub Actions CD Workflow

• Manual trigger: Workflow dispatch with patch/minor/major version selection
• Semantic versioning: Automatic version calculation from latest Git tags
• Multi-architecture builds: Docker buildx for amd64/arm64 platforms
• Registry publishing: Automated push to ghcr.io/graphprotocol/service-quality-oracle
• Release automation: Git tagging and GitHub release creation with changelogs

🔧 chore: Dockerfile Enhancement

• Dynamic versioning: Accept VERSION build argument (defaults to "dev")
• Metadata labels: Version information embedded in Docker image

📚 docs: Comprehensive Documentation

• CD_PROCESS.md: Step-by-step guide for using the CD system
• CHANGELOG.md: Complete project history following Keep a Changelog format
• Release strategies: Examples for single PR vs batched releases

Usage

After merge, releases are triggered manually:

1. Go to Actions → "CD - Build and Release"
2. Click "Run workflow"
3. Select version type (patch/minor/major)
4. Docker images published as ghcr.io/graphprotocol/service-quality-oracle:vX.Y.Z

Future Improvements

• Consider adding release candidate (RC) builds for testing
• Add metrics/monitoring for CD pipeline success rates
• Evaluate cache warming strategies for faster builds
• Consider automated CHANGELOG.md updates in future iterations

Risk Assessment

Low Risk:

• Manual trigger prevents accidental releases
• Existing functionality unchanged - purely additive
• Docker builds tested and validated
• No breaking changes to current deployment process

Notes for Reviewers

• First release will be v0.2.0 (v0.1.0 documented as BigQuery caching release)
• Container registry: Images will be public on ghcr.io (appropriate for open source)
• Workflow permissions: Uses standard GITHUB_TOKEN with minimal required permissions
• Multi-arch builds: May take 5-10 minutes due to cross-compilation
• Documentation: Both user-facing (CD_PROCESS.md) and project history (CHANGELOG.md) included

The CD system provides complete control over release timing while automating the complex build/publish/tag workflow.