Top 30 Scalability Interview Questions and Answers
30 multiple-choice questions on Scalability, drawn from 30 bites out of the 64 tagged Scalability on Tezvyn. Answer them here or read straight down. Every question carries the correct option, why it is correct, and a link to the bite it came from.
30 questions. Pick an answer, or open “Show the answer” to read it.
Answers are graded in your browser. Nothing is saved, and no XP or streak is earned here. The app keeps score.
Question 1 of 30
What problem does cloud computing primarily solve for businesses launching new digital services?
Show the answer
Answer: c · The high upfront costs and risks of owning physical IT infrastructure.
The card states cloud computing was created to solve "huge upfront capital costs, long procurement times, and the risk of buying too much or too little capacity" associated with owning physical servers. While cloud computing can help with IT personnel challenges (A) and high availability (D), these are not its primary foundational purpose.
Read the full bite: Cloud Computing: Renting Someone Else's Computer
Question 2 of 30
What is the fundamental principle guiding Go's design?
Show the answer
Answer: d · Optimizing for practical engineering challenges in large systems.
Go was created to solve practical problems like slow builds and complexity in massive codebases, prioritizing engineering concerns such as maintainability and team productivity. It explicitly avoids being a research language or focusing on novel paradigms.
Read the full bite: Go's Design Philosophy: Engineering Over Novelty
Question 3 of 30
Which scenario specifically demonstrates elasticity rather than just scalability?
Show the answer
Answer: b · An auto-scaling group automatically adds instances during a traffic spike and removes them when it subsides
Elasticity is the automatic, two-way matching of capacity to demand, including scaling back down. Manual additions, code rewrites, and one-time vertical upgrades show scalability but not the automatic contraction that defines elasticity.
Question 4 of 30
For which scenario would a full design system typically be considered an unnecessary investment?
Show the answer
Answer: c · A small team developing an early-stage, rapidly iterating prototype
The card states that for a single, small-scale project or a rapidly iterating early-stage prototype, building a full design system is often overkill due to the initial investment. The other options describe situations where a design system provides significant benefits, such as consistency, efficiency, and shared understanding, making them scenarios where it is recommended.
Read the full bite: Design Systems: Your UI's Single Source of Truth
Question 5 of 30
Which cloud strategy combination best addresses predictable monthly user growth and unpredictable, short-duration traffic spikes?
Show the answer
Answer: b · Implement scalability for the monthly growth and elasticity for the celebrity-induced spikes.
Scalability is ideal for planned, long-term growth, while elasticity efficiently handles unpredictable, short-term demand fluctuations by automatically adjusting resources. Relying only on elasticity for steady growth is inefficient, as it would constantly make small adjustments instead of operating from a stable baseline.
Read the full bite: Cloud Scalability vs. Elasticity: Planned Growth vs. Real-Time Reaction
Question 6 of 30
What is the core advantage of implementing a 'system of systems' for design management?
Show the answer
Answer: c · It provides a framework to balance shared brand foundations with diverse product-specific requirements.
The 'system of systems' approach is designed to balance a central core of shared brand elements with the flexibility for individual product-specific design systems to address unique needs. Option B is incorrect because the card explicitly states that a single monolithic system often fails at scale, and this approach is a departure from that.
Read the full bite: System of Systems: Managing Design Systems at Scale
Question 7 of 30
A users table is sharded by user_id. What is the most efficient way to support frequent logins that look users up by email?
Show the answer
Answer: a · Maintain a secondary email-to-user_id index to resolve the shard in one hop
A secondary mapping from email to user_id lets a login resolve the correct shard directly, avoiding a broadcast. A UNIQUE constraint only enforces uniqueness within a single shard, and scatter-gather wastes resources on every login.
Read the full bite: Shard key impact on uniqueness and cross-shard lookups
Question 8 of 30
When an application scales by adding new servers, how does a cloud load balancer ensure these new servers are utilized?
Show the answer
Answer: b · It detects the new servers via health checks and includes them in its traffic distribution.
The card states that the load balancer "automatically detects these new instances, passes its health checks, and begins routing a share of the incoming HTTP requests to them." This ensures new servers are utilized. Option C is incorrect because the load balancer manages the backend server pool internally, not by updating DNS for individual backend servers.
Read the full bite: Cloud Load Balancer: Your App's Traffic Cop
Question 9 of 30
What is the main advantage of using component variants in a design system?
Show the answer
Answer: b · They allow a single base component to represent many different forms by adjusting its properties.
The card states that variants allow "one core component to serve multiple, predictable purposes" by defining "properties and their possible values." This directly aligns with option B. Option A is incorrect because variants are for managing variations of existing components for repeatable patterns, not for creating new or unique, non-repeatable elements.
Read the full bite: Component Variants: One Component, Many Forms
Question 10 of 30
Which scenario best demonstrates the primary benefit of denormalization?
Show the answer
Answer: d · An analytics dashboard displaying pre-calculated sales trends over time.
Denormalization is ideal for read-heavy systems like analytics dashboards, where pre-calculating or duplicating data significantly speeds up frequent queries. It is generally avoided in write-heavy systems or when strict data integrity and minimal storage are paramount.
Read the full bite: Denormalization: Trading Write Speed for Faster Reads
Question 11 of 30
How should a high-scale headless CMS prevent thundering herds on a viral article while ensuring newly published edits appear immediately?
Show the answer
Answer: a · Use request coalescing at the origin and actively purge CDN edge nodes via surrogate-key invalidation events on every publish
Request coalescing collapses concurrent origin fetches for the same key to prevent thundering herds, while surrogate-key invalidation actively clears edge caches on publish instead of waiting for TTL. Option C is tempting because stale-while-revalidate is a valid resilience tactic, but without explicit invalidation it cannot guarantee that a newly published edit appears immediately.
Read the full bite: Design a highly scalable headless CMS architecture
Question 12 of 30
When architecting a response to a competitor's data-intensive feature, what is the most critical first step to create a defensible performance gap?
Show the answer
Answer: a · Reverse engineer the competitor's likely data or compute bottleneck to find their scaling pain point
The card states that the first step is to reverse engineer the competitor's bottleneck so your distributed design targets their specific scaling pain point. Shipping a feature-complete clone is a common wrong answer because it skips bottleneck analysis and replicates surface functionality without creating an asymmetric data-path advantage.
Read the full bite: How would you out-engineer a competitor's new data-intensive feature?
Question 13 of 30
A Button component has 12 boolean props and 5 enums, creating hundreds of untested states and bundle bloat. Which strategy best addresses root causes without shifting burden entirely to consumers?
Show the answer
Answer: b · Split into single-purpose composable primitives and offer curated high-level patterns for common cases
Composable primitives eliminate invalid prop combinations and enable tree-shaking, while curated high-level components preserve consumer ergonomics. Exhaustive testing of every combination is a superficial fix that fails to reduce the exponential state space.
Read the full bite: Super button versus composition: discuss trade-offs and scalability
Question 14 of 30
Which pattern prevents duplicate coupon reservations when high-volume email jobs are retried due to ESP timeouts?
Show the answer
Answer: d · Atomically reserve the coupon before enqueueing and deduplicate retries with deterministic job IDs
Atomic reservation before enqueueing coupled with deterministic job IDs ensures exactly one code is mapped per user even during at-least-once retries. Generating coupons during SMTP seems safe but burns inventory on network timeouts without guaranteed delivery.
Read the full bite: How would you architect personalized email and coupon delivery at scale?
Question 15 of 30
Which label choice is most likely to cause a damaging cardinality explosion in a metrics system?
Show the answer
Answer: c · Per-request unique request_id
A unique request_id is unbounded, creating a new series per request and exploding cardinality. The other labels have small bounded value sets, so they stay cheap.
Read the full bite: High cardinality in time-series databases
Question 16 of 30
What is the primary goal of conducting software performance testing?
Show the answer
Answer: d · To evaluate the system's stability, responsiveness, and resource utilization under anticipated and extreme user loads.
Performance testing specifically aims to understand how a system behaves under various loads, measuring its stability, responsiveness, and resource usage to find breaking points. Options A, B, and C describe functional testing, security testing, and general bug fixing, respectively, which are distinct from performance evaluation under load.
Read the full bite: Software Performance Testing: How a System Behaves Under Stress
Question 17 of 30
When running two identical HA Prometheus replicas behind a global query layer, what must the query layer do?
Show the answer
Answer: c · Deduplicate overlapping series from the replicas
Identical replicas produce overlapping data, so the query layer must deduplicate to avoid double-counting while still surviving one replica failing. Summing would inflate results; permanently ignoring a replica defeats HA.
Question 18 of 30
What is the primary reason ITCSS orders layers from generic to specific?
Show the answer
Answer: c · So specificity rises with source order, letting later rules override earlier ones predictably
Ordering low-to-high specificity along source order means later, more specific rules cleanly win, avoiding specificity wars and !important. The other options misattribute the benefit to file size, parse speed, or removing classes.
Question 19 of 30
When designing a low-latency entitlement system, how should the hot path handle plan validation and quota checks?
Show the answer
Answer: c · Issue signed edge tokens for plan validation and stream quota usage to an async pipeline
The correct approach separates the user-facing hot path from background policy work by validating signed edge tokens locally and streaming quota events asynchronously, keeping latency under five milliseconds. A synchronous database lookup per request is a common red flag because it couples the hot path to a remote dependency and creates a scaling bottleneck.
Read the full bite: Propose a scalable entitlement architecture for complex rules
Question 20 of 30
Why is embedding the full list of liking user IDs inside each post document a poor choice for a popular social platform?
Show the answer
Answer: b · The liker array is unbounded, so it can exceed the document size limit and forces rewriting the whole document on each like
Unbounded growth collides with the document size cap and causes costly full-document rewrites and contention per like. Document stores do support arrays and offer lookups, so the other options are false.
Read the full bite: Embed or reference likes in a document database?
Question 21 of 30
What most reliably keeps thousands of programmatically generated landing pages from being flagged as thin content?
Show the answer
Answer: c · Ensuring each page is backed by genuinely distinct, substantive data
Thin-content penalties target near-duplicate pages differing only by a token; genuinely distinct per-page data gives each page unique value. Keyword stuffing the template or sitemap mechanics do nothing about the underlying sameness.
Read the full bite: Design a programmatic SEO landing-page system
Question 22 of 30
Which architectural property of Cassandra most directly removes a write bottleneck for a high-write user profile service?
Show the answer
Answer: a · A masterless peer-to-peer design where any replica can accept writes, with data spread by consistent hashing
Cassandra's leaderless ring lets any replica accept writes and spreads them via consistent hashing, eliminating a single write bottleneck. A single primary, synchronous all-replica writes, and joins would each constrain rather than help write throughput.
Read the full bite: Why fit Cassandra to a high-read, high-write workload?
Question 23 of 30
When designing a hybrid feed system that switches between fan-out-on-write and fan-out-on-read, how should you determine the follower threshold T?
Show the answer
Answer: a · Derive T from operational write throughput limits and the system's post rate
The card specifies that T should be calculated from operational limits like acceptable write throughput divided by post rate, not guessed or based on averages. Option C confuses average distribution with peak write capacity, while D treats the threshold as an arbitrary constant rather than a derived operational bound.
Read the full bite: Compare fan-out-on-write vs fan-out-on-read for an activity feed
Question 24 of 30
A team following Twelve-Factor Factor III stores configuration in environment variables. What operational benefit does this provide for CI/CD and scalability?
Show the answer
Answer: a · One build artifact can be promoted across environments and new instances start with the correct context immediately.
Storing config in environment variables keeps the codebase identical across stages, enabling a single artifact to be promoted through CI/CD and allowing new instances to read the correct settings at startup for horizontal scaling. The distractor about keeping only secrets in env vars is wrong because Factor III applies to all deployment-specific configuration, not just sensitive data.
Read the full bite: What is Twelve-Factor's config recommendation for CI/CD and scalability?
Question 25 of 30
To support millions of scheduled personalized notifications without overwhelming downstream providers, which architectural pattern should you prefer?
Show the answer
Answer: a · Stream events into a delay queue and process due jobs with idempotent, rate-limited workers
The card recommends decoupling ingestion from execution via an event stream and distributed delay queue, using idempotent workers that respect rate limits. Option D is tempting because database polling is a common pattern, but the card explicitly flags it as a red flag that creates hot shards and lacks ordering guarantees.
Read the full bite: Design a real-time personalized notification trigger system
Question 26 of 30
How should the pipeline distinguish handling of a 4xx invalid-address error from a 5xx ESP timeout?
Show the answer
Answer: b · Route 4xx errors to a dead-letter queue and retry 5xx errors with exponential backoff and jitter
4xx errors indicate permanent client failures like invalid addresses and should move to a dead-letter queue rather than consuming retry budget, while 5xx timeouts are transient and warrant exponential backoff with jitter. Retrying 4xx wastes throughput and risks reputation damage, whereas dropping 5xx loses valid emails.
Read the full bite: Build a system to send 1M personalized emails in 2 hours
Question 27 of 30
What are the two primary benefits of database replication, and how does it differ from sharding?
Show the answer
Answer: d · It keeps full copies on multiple nodes, giving high availability via failover and better read scalability, unlike sharding which partitions data
Replication maintains full copies for failover and spread-out reads, whereas sharding partitions data across nodes. It does not chiefly boost write throughput, and it complements rather than replaces backups.
Read the full bite: What is database replication and why use it?
Question 28 of 30
Why might a team shard a database rather than continue vertically scaling a single server?
Show the answer
Answer: b · Vertical scaling hits a hardware ceiling, grows disproportionately costly, and is a single point of failure, while sharding distributes data and write load across nodes
Sharding spreads data and write load across commodity nodes, sidestepping the cost ceiling and single point of failure of one big server. Copying the whole dataset describes replication, and sharding requires a deliberate shard key.
Read the full bite: What is sharding and why shard over vertical scaling?
Question 29 of 30
What is the central trade-off between range-based and hash-based sharding?
Show the answer
Answer: a · Range sharding enables efficient range scans but risks hot spots on monotonic keys; hash sharding spreads load evenly but makes range queries inefficient
Range sharding keeps ordered keys together for fast range scans but concentrates monotonic writes on one shard, while hashing distributes load evenly at the cost of efficient range queries. The other options invert these properties.
Read the full bite: Range-based vs hash-based sharding trade-offs?
Question 30 of 30
Across five stateless Express instances behind a load balancer, what is the main operational cost of choosing session-based auth over JWTs?
Show the answer
Answer: a · You need a shared session store so any instance can resolve the session
Session state must be reachable by whichever instance handles a request, so a shared store like Redis is required. JWTs avoid that by being self-contained and locally verifiable; sessions absolutely can scale, just with shared storage.
Read the full bite: Session-based versus token-based authentication
Could you explain these out loud?
That is what an interview actually tests. Tezvyn gives you questions like these with what the interviewer is really checking, the answer that lands, and the mistake that ends the conversation, in the four minutes before your next meeting.
The iPhone app is on the way
We are building it. Until it lands, nothing here is held back from you: every interview card, your saved cards, streaks and the job board all work in Safari, plus hundreds of free practice quizzes of thirty questions each. Sign in and it all carries over to the app the day it arrives.
Want it as an icon? Tap Share at the bottom of Safari, then Add to Home Screen. It opens full screen and the cards you have read stay available offline.