supabase-reliability-patterns

# Supabase Reliability Patterns ## Overview Production Supabase apps need six reliability layers: **circuit breakers** (stop calling Supabase when it's down to prevent cascading failures), **offline queue** (buffer writes when the network is unavailable and replay when reconnected), **graceful degradation** (serve cached or fallback data during outages), **health checks** (detect Supabase availability before routing traffic), **retry with exponential backoff** (handle transient errors without overwhelming the service), and **dual-write** (write critical data to both Supabase and a backup store). All patterns use real `createClient` from `@supabase/supabase-js`. ## Prerequisites - `@supabase/supabase-js` v2+ installed - TypeScript project with Supabase client configured - For offline queue: browser environment with IndexedDB or server with Redis - For dual-write: secondary data store (Redis, DynamoDB, or local SQLite) ## Step 1 — Circuit Breaker and Retry with Exponential Backoff A circuit breaker tracks failures per Supabase service (database, auth, storage) and stops making calls when a threshold is exceeded. Combined with retry logic, it prevents both cascading failures and unnecessary retries during extended outages. ### Circuit Breaker Implementation ```typescript // lib/circuit-breaker.ts import { createClient } from '@supabase/supabase-js' import type { Database } from './database.types' type CircuitState = 'CLOSED' | 'OPEN' | 'HALF_OPEN' interface CircuitBrea...