Massive Traffic
- Massive Traffic - Practice Exercises
- Exercise 1: Load Shedding Strategy
- Exercise 2: Rate Limiting Design
- Exercise 3: Backpressure with Channels
- Exercise 4: Cache-Aside for Hot Reads
- Exercise 5: Read Replica Routing
- Exercise 6: Idempotent Commands
- Exercise 7: Circuit Breakers
- Exercise 8: Queue Consumer Scaling
- Exercise 9: Observability Signals
- Exercise 10: Graceful Degradation
Massive Traffic - Practice Exercises
Short set of practice prompts for high-throughput scenarios. Additional topic files will expand this area.
---
Exercise 1: Load Shedding Strategy
Q: When should you drop requests and how do you signal it to clients?
A: Use 429/503 with retry hints when the system is saturated.
---
Exercise 2: Rate Limiting Design
Q: Compare token bucket vs sliding window.
A: Token bucket smooths bursts; sliding window gives tighter limits.
---
Exercise 3: Backpressure with Channels
Q: Use a bounded channel to prevent overload.
A: Bounded channels block producers when full, protecting downstream.
---
Exercise 4: Cache-Aside for Hot Reads
Q: Implement a cache-aside pattern for quote lookups.
A: Check cache first, fall back to DB, then populate cache.
---
Exercise 5: Read Replica Routing
Q: Route read-only queries to replicas.
A: Use separate read/write connections or a query router.
---
Exercise 6: Idempotent Commands
Q: Make order placement idempotent during retries.
A: Use idempotency keys and dedupe storage.
---
Exercise 7: Circuit Breakers
Q: Protect a downstream price service.
A: Trip the breaker after repeated failures and serve cached data.
---
Exercise 8: Queue Consumer Scaling
Q: Scale consumers based on backlog.
A: Use metrics on queue depth to scale horizontally.
---
Exercise 9: Observability Signals
Q: Which metrics matter most under massive traffic?
A: Latency percentiles, error rate, queue depth, and saturation.
---
Exercise 10: Graceful Degradation
Q: How do you degrade features under load?
A: Disable non-critical endpoints and serve cached responses.