SLA, SLO, and SLI definitions in HLD - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding SLA, SLO, and SLI

Which of the following best describes the relationship between SLA, SLO, and SLI?

A SLI is the agreement, SLA is the measurement, and SLO is the penalty.

B SLA is the metric, SLO is the agreement, and SLI is the penalty for not meeting the agreement.

C SLO is the service metric, SLA is the target, and SLI is the customer's feedback.

D SLI measures a service metric, SLO sets a target for that metric, and SLA is the formal agreement based on the SLO.

💡 Hint

Think about which term is a measurement, which is a goal, and which is a contract.

An SLI (Service Level Indicator) is a specific metric that measures service quality. An SLO (Service Level Objective) is the target value for that metric. An SLA (Service Level Agreement) is the formal contract that includes the SLO and consequences if the SLO is not met.

🧠 Conceptual

intermediate

Identifying an SLI example

Which of the following is an example of a Service Level Indicator (SLI)?

A The percentage of successful HTTP requests over a period.

B 99.9% uptime guaranteed in the contract.

C A penalty clause for downtime exceeding 0.1%.

D A customer agreement to pay for premium support.

💡 Hint

SLI is a measurable metric, not a contract or penalty.

SLI is a metric that measures service quality, such as the percentage of successful HTTP requests. The other options describe agreements or penalties, not metrics.

❓ Architecture

advanced

Designing Monitoring for SLI Collection

You need to design a system to collect SLIs for a web service's availability and latency. Which architectural component is essential to ensure accurate and scalable SLI collection?

A A centralized logging system that collects all request logs and computes SLIs offline daily.

B Distributed agents on each server that collect metrics in real-time and push to a metrics aggregator.

C A manual process where engineers review logs weekly to estimate SLIs.

D A client-side script that reports user satisfaction surveys.

💡 Hint

Consider real-time, scalable, and accurate metric collection.

Distributed agents collecting metrics in real-time and pushing to a central aggregator allow accurate, scalable, and timely SLI measurement. Centralized logging with offline processing is slower and less scalable. Manual review is not scalable or timely. Client-side surveys measure user experience but are not direct SLIs.

❓ scaling

advanced

Scaling SLA Enforcement for a Global Service

Your service has SLAs with different customers worldwide, each with unique SLOs. What is the best approach to scale SLA enforcement and reporting?

A Maintain a single global SLA enforcement system that applies the strictest SLO to all customers.

B Outsource SLA enforcement to a third-party vendor without integration to your monitoring.

C Implement a multi-tenant SLA enforcement system that tracks each customer's SLOs separately and generates customized reports.

D Ignore individual SLAs and focus only on average service metrics.

💡 Hint

Think about customization and scalability for multiple customers.

A multi-tenant SLA enforcement system that tracks each customer's SLOs separately allows accurate enforcement and reporting per contract. Applying the strictest SLO globally is unfair and inaccurate. Outsourcing without integration loses control. Ignoring individual SLAs risks contract violations.

❓ tradeoff

expert

Tradeoff Between SLI Granularity and System Overhead

You want to increase the granularity of SLIs by measuring latency per API endpoint every second. What is a key tradeoff to consider?

A Higher granularity improves accuracy but increases monitoring system overhead and storage costs.

B Higher granularity reduces accuracy but lowers system overhead.

C Higher granularity eliminates the need for SLAs.

D Higher granularity guarantees zero downtime.

💡 Hint

Think about the balance between detail and resource use.

Increasing granularity improves accuracy and insight but also increases the load on monitoring infrastructure and storage needs. It does not reduce accuracy or eliminate SLAs. It cannot guarantee zero downtime.