Skip to content

Live

Home › Explanation › Continuous Delivery › Core Concepts › Live

Introduction

"Live" is the destination where released software becomes available to consumers.

What "Live" means depends on the module type: a production environment for services, or an artifact registry for libraries.

Live Legend

Definition

Live: The destination where released software serves consumers.

Live takes different forms depending on the deployable module type:

Module Type Live Destination Consumers
Services Production environment End users
Libraries Package registry (NuGet, npm, PyPI) Developers
CLI tools Package manager or release page Users/automation
Container images Container registry (Docker Hub, GHCR) Orchestrators
Documentation Published site (GitHub Pages) Readers

For Deployed Services

Services run continuously in production environments serving real users.

Characteristics:

  • Real user traffic and business data
  • High availability requirements
  • Continuous monitoring and alerting
  • Incident response procedures

Key Capabilities:

  • Monitoring: Application performance, error rates, resource utilization
  • Observability: Distributed tracing, log aggregation, metrics dashboards
  • Feature Flags: Runtime control over feature availability
  • Rollback: Redeploy previous version if issues detected

CD Model Stages:

  • Stage 10 (Production Deployment): Deploy to production
  • Stage 11 (Live): Monitor health and performance
  • Stage 12 (Release Toggling): Enable/disable features via flags

For Published Artifacts

Libraries and tools are published to registries where consumers pull specific versions.

Characteristics:

  • Immutable versioned releases
  • Multiple versions available simultaneously
  • Consumers control which version they use
  • Deprecation process for old versions

Key Capabilities:

  • Semantic Versioning: Signal compatibility (breaking vs non-breaking)
  • Download Metrics: Track adoption and usage
  • Deprecation Notices: Warn consumers of outdated versions
  • Security Advisories: Alert consumers to vulnerabilities

CD Model Stages:

  • Stage 10 (Production Deployment): Publish to registry
  • Stage 11 (Live): Track downloads and adoption
  • Stage 12 (Release Toggling): Mark versions as latest/deprecated

Feedback Loops

Live environments generate feedback that influences future development:

Live Environment
    │
    ├─► Monitoring alerts ─► Investigation ─► Bug fixes to Trunk
    ├─► Performance data ─► Optimization ─► Improvements to Trunk
    ├─► User behavior ─► Product insights ─► Features to Trunk
    └─► Incidents ─► Post-mortems ─► Process improvements

For Services:

  • Error rates trigger alerts
  • Performance degradation initiates optimization
  • Feature flag data informs product decisions
  • Incidents result in fixes committed to trunk

For Artifacts:

  • Download trends inform prioritization
  • Issue reports drive bug fixes
  • Consumer feedback shapes roadmap
  • Security scans trigger patch releases

Rollback Strategies

When issues occur in live, rollback provides recovery:

Services:

Strategy How Speed
Redeploy previous Deploy last known good version Minutes
Feature flag Disable problematic feature Seconds
Blue-green switch Route traffic to previous deployment Seconds
Database rollback Restore data (if applicable) Varies

Artifacts:

Strategy How Consumer Impact
Yank/unpublish Remove bad version from registry Breaks builds using it
Patch release Publish fixed version Consumers must upgrade
Pin recommendation Advise pinning to safe version Manual consumer action

Next Steps

References

Tutorials | How-to Guides | Explanation | Reference

You are here: Explanation — understanding-oriented discussion that clarifies concepts.