Practice : Deployment Freeze Windows
Purpose and Strategic Importance
Deployment Freeze Windows are scheduled periods during which no production deployments or high-risk changes are allowed. These windows are typically aligned with business-critical events, holidays, or support limitations - helping to reduce risk, preserve stability, and maintain service continuity.
Used thoughtfully, freeze windows can balance engineering velocity with operational safety. They provide predictability and reduce the likelihood of incidents during high-risk periods, supporting customer trust and internal well-being.
Description of the Practice
- A defined calendar or schedule outlines when deployments are paused.
- Freeze windows apply to high-risk or production-impacting changes.
- Exceptions are clearly defined and follow a formal approval process.
- Monitoring, alerting, and rollback capabilities remain active during freezes.
- Teams are notified well in advance and adjust delivery plans accordingly.
How to Practise It (Playbook)
1. Getting Started
- Define a freeze window calendar aligned to business events (e.g. holidays, peak usage periods).
- Communicate freeze windows early through engineering channels and rituals.
- Clarify what types of changes are restricted (e.g. production deployments, infra updates).
- Establish a clear exception process - include stakeholders and approval paths.
2. Scaling and Maturing
- Use CI/CD tooling to enforce freeze windows automatically (e.g. pipeline schedules, deployment locks).
- Define risk-based tiers - not all changes require the same freeze level.
- Ensure observability, on-call coverage, and rollback tools remain active throughout.
- Include freeze planning in sprint cycles and release planning cadences.
- Regularly review freeze effectiveness and make adjustments based on incident trends.
3. Team Behaviours to Encourage
- Plan around freezes - avoid last-minute deployments before critical periods.
- Collaborate across product and engineering to balance urgency and safety.
- Treat freezes as a protective measure, not a blocker - focus on quality and stability.
- Communicate early, especially when freeze windows overlap with delivery milestones.
4. Watch Out For…
- Overly broad or unnecessary freezes that stall delivery without justification.
- Manual enforcement - automate wherever possible.
- Rushing risky changes before a freeze begins.
- Inflexible policies that prevent safe and necessary updates (e.g. security patches).
5. Signals of Success
- Teams proactively plan and deliver around freeze windows.
- Incidents during freeze periods decrease over time.
- Freeze policies are well-understood, not burdensome.
- Approvals for exceptions are fast, justified, and aligned with risk.
- Teams use freeze windows for hardening, observability, and reliability improvements.