Pro-Owner perspective: This document frames your systems as a technical estate — an asset to be stewarded, documented, and bequeathed. Treat these steps as craftsmanship: protect the continuity, auditability, and transferability of your digital legacy.
What it is
A structured system for scheduling infrastructure changes (deployments, config updates, patches) within pre-approved maintenance windows, with mandatory freeze periods during high-risk business periods. Changes are risk-tiered: Standard (tested, scheduled), Expedited (urgent, higher approval), Emergency (incident response, post-facto approval).
Freeze periods are non-negotiable blackout windows (e.g., Black Friday week, year-end close) where only emergency changes are allowed. Change success/failure rates are tracked by tier to identify risky change patterns.
Why it matters
Uncoordinated changes create incident clusters ("we deployed three things and now everything's broken—which one?"). Change windows batch changes for easier rollback, concentrate monitoring attention, and allow pre-communication to stakeholders.
Freeze periods protect revenue-critical periods from change-induced outages. Without freezes, you risk deploying a bug-fix-that-breaks-checkout during your biggest sales week.
How we do it
- Change tiers:
- Standard: Routine changes (patches, minor config updates). Scheduled 48 hours ahead. Approval: tech lead. Window: Tuesday/Thursday 10pm-2am.
- Expedited: Urgent fixes (security patches, performance issues). Scheduled 24 hours ahead. Approval: director. Window: Any weekday evening.
- Emergency: Incident response only. No advance scheduling. Approval: incident commander. Window: Immediate. Post-facto documentation required.
- Change calendar: Annual calendar showing:
- Regular maintenance windows (weekly Tuesday/Thursday)
- Extended windows (monthly Saturday for major upgrades)
- Freeze periods (Thanksgiving week, year-end, peak season)
- Change request process:
- Submit change request: what, why, risk assessment, rollback plan, testing evidence.
- Approval based on tier and window availability.
- Pre-communication to stakeholders (internal teams, customers if user-facing).
- Post-change validation: success metrics, incidents triggered (if any), lessons learned.
- Freeze period management:
- Freeze periods announced 90 days ahead.
- Override process: Emergency changes only, with executive approval and documented business justification.
What you receive
- Change calendar: Annual view of windows, freeze periods, scheduled changes.
- Change approval workflows: Forms, approval chains, SLAs by tier.
- Change metrics: Success rate by tier, incident correlation, time-to-rollback.
- Freeze override log: Emergency changes during freezes, with justification and outcomes.
All changes logged in change management system (Jira, ServiceNow, Linear) for audit trail.
Evidence
Interactive change heatmap:
- Calendar view showing risk tiers (color-coded: green=standard, yellow=expedited, red=emergency).
- Freeze periods overlaid (grey-out, no changes allowed).
- Toggle seasonality presets (retail, SaaS, healthcare) to see industry-typical freeze patterns.
- Click date to see change request template for that tier/window.
Download change calendar package (templates + approval workflows + metrics dashboard): [Link]
Failure modes & guardrails
Failure mode: Too many emergency changes
Guardrail: If emergency changes exceed 10% of total changes in a quarter, mandate root cause analysis and process improvement.
Failure mode: Freeze periods ignored
Guardrail: Freeze overrides require CEO approval (non-delegable). Every override creates a post-mortem report.
Failure mode: Change windows too narrow (batching failures)
Guardrail: If change success rate < 90%, expand windows or add additional window slots.
Failure mode: No rollback plan
Guardrail: Change request rejected if rollback plan missing or untested. No exceptions.