Skill: rollout-strategy

Skill: rollout-strategy

Purpose

Select and design the right rollout strategy for a release — balancing speed, risk, and user impact.

When to Use

• During release planning
• When releasing to a large or sensitive user base
• When the feature has high technical or business risk

Framework: Rollout Strategy Selection

Step 1 — Understand the Context

Ask for:

• User base size and type (internal, B2B, B2C)
• Feature risk level (high / medium / low)
• Rollback complexity (easy / hard / very hard)
• Monitoring capability (robust / basic / none)

Step 2 — Strategy Options

Option A: Big Bang

Description: Deploy to 100% of users at once. Best for: Low-risk changes, internal tools, hotfixes, small user bases. Risk: High — if something breaks, everyone is affected. Rollback: Redeploy previous version.

Option B: Phased Rollout (Percentage-Based)

Description: Release to 5% → 25% → 50% → 100% over days/weeks. Best for: B2C products, large user bases, medium-risk features. Risk: Medium — issues caught early affect small % of users. Rollback: Reduce percentage to 0%.

Option C: Canary Release

Description: Release to a small "canary" group (e.g., internal employees or power users) before general availability. Best for: High-risk features, new infrastructure, major architecture changes. Risk: Low — issues are caught before wide exposure. Rollback: Simple — revert canary group.

Option D: Feature Flags

Description: Deploy code to all users but hide behind a flag. Enable selectively. Best for: When you want zero-downtime deploys and gradual enablement. Risk: Low — can toggle off instantly. Rollback: Disable flag (seconds).

Option E: Blue/Green Deployment

Description: Run two identical environments; switch traffic from Blue (old) to Green (new). Best for: Zero-downtime deployments, critical infrastructure. Risk: Low — instant rollback by switching back to Blue. Rollback: Redirect traffic to Blue (seconds).

Step 3 — Strategy Recommendation Matrix

Context	Recommended Strategy
Low risk, small user base	Big Bang
Medium risk, large B2C	Phased Rollout
High risk, unknown stability	Canary + Feature Flags
Critical infrastructure	Blue/Green
Gradual feature adoption	Feature Flags

Step 4 — Rollout Plan

Based on selected strategy, produce:

Phased Rollout Example:

Phase	% Users	Duration	Success Criteria	Rollback Trigger
Phase 1	5%	24 hours	Error rate < 0.5%, p95 latency < 300ms	Error rate > 1%
Phase 2	25%	48 hours	Same + no P0 bugs	Any P0 bug
Phase 3	100%	—	Stable for 72 hours	—

Output Format

1. Strategy recommendation with rationale
2. Rollout plan table with success criteria and rollback triggers
3. Monitoring checklist for launch day
4. Suggested next command: /launch-checklist or /assess-readiness