Skip to main content

Advanced Recommendations for Secure Deployment

This document complements the base practices of Chapter 11 - Secure Deployment with technical and organisational recommendations for contexts of greater maturity, increasing the resilience, traceability and auditability of deliveries to production.

The practices described here are not mandatory, but highly recommended for applications classified as L3 or in regulated environments.

1. Deployment Validated on the Basis of Observability​

  • Use of dashboards with readiness metrics for automatic gating;
  • Configuration of canary release with automatic promotion and rollback.

2. Reversibility Validated in the Pipeline​

  • Execution of rollback in staging before each release;
  • Scripts tested and documented as part of the build.

3. Signed Release Contract​

  • release-contract.yaml file containing:
    • Versions, owners, validations, exceptions, rollback;
  • Digital signature or hash for later validation.

4. Deployment with Provenance Verification​

  • Validation of SBOM, signature, provenance and build agent;
  • Requirement for authorised pipelines.

5. Formalised "Break-Glass" Policy​

  • Approved process for emergency deployments:
    • Mandatory audit;
    • Dual approval (product + security);
    • Planned post-event rollback.

6. Post-Production Security Verification​

  • Execution of runtime-aware tests:
    • HTTP headers, ports, endpoints;
    • Hardening validation.

7. Dashboards with Deployment Traceability​

  • Visualisation per release:
    • Date, owners, validations, findings, rollback and SLOs.

8. Multi-Factor Approval of Critical Deployments​

  • Approval by security + product + operations;
  • Based on a checklist and technical evidence.

9. Reinforced Immutability at Runtime​

  • Prohibition of post-deployment modifications;
  • Rejection of manual patches or “hotfixes” outside the build cycle.

10. Environment Integrity Verification​

  • Checksum, versions, permissions, configurations and environment variables;
  • Validation before and after deployment.

11. Release Freezing with Time-Based Revalidation​

  • Automatic freeze of releases that remain more than X days without promotion;
  • Mandatory re-execution of security validations before deployment;
  • Avoids the promotion of obsolete artefacts without revalidation.

12. Secure Configuration Management in Deployment​

  • Validation of the integrity of secrets, feature toggles and config maps;
  • Policy enforcement via OPA/Rego or Kyverno;
  • Approval of sensitive changes via controlled PRs.

13. Automatic Lockdown after Critical Deployment​

  • Application of a temporary lockdown after critical deployments:
    • Deactivation of experimental toggles;
    • Restriction of administrative access;
    • Reinforced logging until stabilisation.

✅ Conclusion​

These advanced practices extend the chapter's base model, covering scenarios of high operational demand, including:

  • Management of old artefacts and expiry of validations;
  • Security and traceability of configuration in real time;
  • Lockdown and proactive containment of risk in production.

📌 Their application is especially relevant in contexts of high availability, high reputational impact or regulatory requirements.

💡 Deployment maturity does not depend only on the tool used, but on the ability to take secure, justified and auditable decisions about what goes into production.