A DevOps engineer updated a microservice's environment configuration: the CACHE_TTL environment variable for a Redis-backed product catalog service was changed from 300 seconds to 3600 seconds to reduce database load. You need to scout for potential regressions caused by this config change. Write your findings to report.md.

=============== FILE: inputs/catalog-service.md ===============

Product Catalog Service

Configuration Change

• CACHE_TTL: 300 → 3600 seconds (1 minute → 1 hour)
• Applied to: all product cache entries in Redis

Key Components

• services/catalog.py — product lookup, caches results with CACHE_TTL
• services/inventory.py — inventory checks, reads from catalog cache
• services/pricing.py — price lookups, may read stale catalog data from cache
• api/routes/products.py — product listing and detail endpoints
• api/routes/checkout.py — checkout flow reads product prices
• middleware/auth.py — admin routes for cache invalidation require ADMIN role
• tasks/cache_warmer.py — background task that pre-warms cache on startup

Known Behaviors

• Price updates in the DB would previously appear within 5 minutes
• With new TTL, stale prices could persist for up to 1 hour
• Cache warmer runs every 5 minutes but respects existing TTL (won't overwrite)
• Admin cache-invalidation endpoint requires ADMIN role JWT claim =============== END FILE ===============