14.6 Caching & Idempotency Key Stores

1. What It Is in Plain Words

Caching stores copies of frequently accessed or slow-to-compute data in fast, in-memory memory stores (like Redis) to accelerate read queries and reduce database load.

---

2. How This Project Uses It

In backend/src/aggregation/counters.py:

• Write-Through Ingestion Cache: Redis atomic INCRBY counters incremented as telemetry arrives, powering real-time dashboard meters.
• Idempotency Cache: Redis key-value store caching idempotency_key hashes with a 24-hour TTL to short-circuit database lookups.

---

3. 30-Second Interview Answer

"We implement a Write-Through caching layer in Redis. Inbound idempotency keys are cached with a 24-hour TTL, allowing duplicate retries to be acknowledged in under 1 millisecond without touching PostgreSQL. Real-time consumption meters use Redis atomic INCRBY counters, providing instant dashboard updates while PostgreSQL remains the durable source of truth."