IntermediateScenario
10 min
Storage IOPS Throttling
StorageCloudPerformance
Advertisement
Interview Question
An application shows sudden latency spikes due to cloud storage IOPS limits being hit. How do you confirm and fix?
Key Points to Cover
- Measure disk latency/IOPS metrics from cloud provider
- Confirm correlation between QPS and throttling events
- Redistribute load or shard data across volumes
- Upgrade to higher IOPS tier or provisioned IOPS
- Introduce caching for hot paths
Evaluation Rubric
Confirms IOPS bottleneck with metrics30% weight
Correlates with workload patterns30% weight
Suggests scaling/caching strategies20% weight
Proposes monitoring and tiering20% weight
Hints
- 💡Burst balance exhaustion is common on gp2 volumes.
Potential Follow-up Questions
- ❓When to use provisioned IOPS?
- ❓What caching layers help?
Advertisement