Alerting thresholds in HLD - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Alerting Thresholds in Monitoring Systems

Which of the following best describes the purpose of setting alerting thresholds in a system monitoring context?

A To define specific limits that trigger notifications when system metrics exceed or fall below normal ranges

B To automatically fix system errors without human intervention

C To store historical data for long-term analysis without triggering any alerts

D To increase the system's processing speed by adjusting resource allocation dynamically

💡 Hint

Think about why monitoring tools notify engineers about unusual system behavior.

Alerting thresholds are set to notify when metrics go beyond expected values, helping teams respond quickly to issues.

❓ Architecture

intermediate

Designing a Threshold-Based Alerting Component

In a high-level design for an alerting system, which component is primarily responsible for evaluating if a metric breaches its threshold?

A Data Collector

B Data Storage

C Threshold Evaluator

D Notification Dispatcher

💡 Hint

Consider which part checks if values are too high or too low.

The Threshold Evaluator compares incoming metric values against predefined limits to decide if an alert should be triggered.

❓ scaling

advanced

Scaling Alerting Threshold Evaluations for Large Systems

When scaling an alerting system to handle millions of metrics per second, which approach best ensures timely threshold evaluations?

A Centralize all metric evaluations on a single powerful server

B Distribute threshold evaluations across multiple nodes close to data sources

C Disable threshold evaluations to reduce load and rely on manual checks

D Store all metrics first and evaluate thresholds in batch once per day

💡 Hint

Think about reducing latency and avoiding bottlenecks.

Distributing evaluations near data sources reduces latency and balances load, enabling timely alerts at scale.

❓ tradeoff

advanced

Tradeoffs in Setting Alerting Thresholds

What is a common tradeoff when setting very sensitive (low) alerting thresholds in a monitoring system?

A Increased false positives leading to alert fatigue

B Missing critical system failures due to high thresholds

C Reduced system performance due to heavy data storage

D Longer alert delivery times due to batching

💡 Hint

Consider what happens if alerts trigger too often for minor issues.

Very sensitive thresholds cause many alerts, including false alarms, which can overwhelm responders and reduce effectiveness.

❓ estimation

expert

Estimating Alert Volume for Threshold Settings

A system monitors 10,000 metrics, each checked every 10 seconds. If 0.5% of checks breach thresholds on average, estimate how many alerts are generated per hour.

A 3,000 alerts per hour

B 30,000 alerts per hour

C 1,800 alerts per hour

D 18,000 alerts per hour

💡 Hint

Calculate total checks per hour, then apply breach percentage.

Each metric is checked 6 times per minute (60/10), so 10,000 * 6 * 60 = 3,600,000 checks per hour. 0.5% of 3,600,000 is 18,000 alerts.