Graceful Degradation with Fallbacks

When primary services fail, don't just show errors. Provide degraded but functional experiences. Fallback patterns define what to do when the ideal path fails.

The Fallback Hierarchy

Design fallback chains from best to worst experience. Each step trades quality for reliability:

Primary: Best quality, might fail
Alternative service: Good quality, more reliable
Cached data: Stale but instant
Simplified version: Reduced features, always works
Static default: Minimal but functional

Four Fallback Strategies

💾

Cached Response

Return previously cached data. Stale is better than nothing.

Best for: Content, recommendations, search results

🔄

Alternative Service

Switch to backup service. Lower quality or cost-effective option.

Best for: AI models (GPT-4 → GPT-3.5), APIs with fallbacks

⚡

Simplified Version

Reduce features or complexity. Basic functionality that always works.

Best for: Search (vector → keyword), recommendations (ML → popular)

📋

Static Default

Hardcoded fallback. Minimal but guaranteed to work.

Best for: Error messages, placeholder content, empty states

Interactive: Fallback Chain Simulator

Select a service scenario and simulate failures to see the fallback chain in action:

Select Service Scenario

🌐Translation Service

Primary Service

Call GPT-4 for high-quality translation

🔄 Fallback 1: Alternative

Try GPT-3.5 (faster, cheaper, slightly lower quality)

💾 Fallback 2: Cached

Return cached translation if available

📋 Fallback 3: Default

Return original text with note "Translation unavailable"

💡

Design Principle

Users rarely notice graceful degradation, but they always notice broken experiences. A slightly slower or lower-quality result that works beats a perfect result that never arrives. Design fallback chains that prioritize availability over perfection.

Error Recovery Strategies

Your Progress