Architecturearchitecture
API design for system design interviews
API design for system design interviews. Resources, verbs, pagination, versioning, and a shortener or feed you can write in five minutes.
All categories
Recent posts
Every published post, newest first. Same list as the feed, without picking a subject.
API gateway for system design interviews
API gateway for system design interviews. The jobs the hop is good at, what must stay behind it, and a chat or feed that needs one.
- Consensus and coordinationavailability
Availability in system design interviews
Availability in system design interviews. What a nine hides, planned versus unplanned work, independent redundancy, and the user-visible error budget.
- Traffic and reliabilitymessaging
Backpressure in system design interviews
Backpressure in system design interviews. Slow consumers, bounded queues, drop versus block, and how you apply it across a network hop.
- Caching and storagestorage
Blob storage for system design interviews
Blob storage for system design interviews. Put bytes in an object store, keep the key in a row, and treat the pointer as a separate consistency problem.
- Caching and storagecaching
Caching for system design interviews
Caching for system design interviews. What a cache may get wrong, write policies in one table, and invalidation you can run.
- Observability and opsestimation
Capacity planning for system design interviews
Capacity planning for system design interviews. QPS, storage, machine count, peak versus average, and the headroom you can defend out loud.
CDN for system design interviews
CDN for system design interviews. Edge versus origin, what you may cache, push versus pull, and a hot video or a hot short link.
- Traffic and reliabilityavailability
Circuit breaker pattern for interviews
Circuit breaker pattern for interviews. Closed, open, and half-open, what trips the breaker, honest fallback, and where it lives next to retries.
- Architecturearchitecture
Clean architecture for system design interviews
Clean architecture for system design interviews. Dependencies that point inward, a payment cut this way, and when a simpler package layout is enough.
- Consensus and coordinationconsistency
Consensus in system design interviews
Consensus in system design interviews. Agreement, validity, and termination in plain words, why a majority works, and where it sits in a drawing.
- Architecturearchitecture
CQRS and event sourcing for interviews
CQRS and event sourcing for interviews. Separate write and read models, optional event store, and the lag a read model is allowed.
Database indexing for system design interviews
Database indexing for system design interviews. What a B-tree buys you, why column order matters, and the write cost you name before you add another index.
Database replication for system design interviews
Database replication for system design interviews. Why a second copy exists, sync versus async, what a reader may see, and a failover that avoids split brain.
Design chat system for the interview
Design chat system for the interview. Open connections, the path of one 1:1 message, group fan-out, history, and presence.
Design distributed cache for the interview
Design distributed cache for the interview. Placement, eviction, a stampede, a stale write, and the node that melts on a hot key.
- Observability and opsarchitecture
Design job scheduler for the interview
Design job scheduler for the interview. Jobs, schedules, and workers, plus catch-up after a missed window and why exactly-once will not hold.
- Observability and opsobservability
Design logging system for the interview
Design logging system for the interview. Agents, a buffer, and a searchable store, plus how logs differ from metrics when a spike arrives.
- Observability and opsobservability
Design metrics system for the interview
Design metrics system for the interview. Emit, transport, store, and query time series, then keep cardinality and alerting off the dashboard path.
- Worked designsscalability
Design notification system for interviews
Design notification system for interviews. Preferences, fan-out, quiet hours, and push, email, and SMS as separate workers.
- Worked designsscalability
Design payment system for the interview
Design payment system for the interview. The ledger, idempotent charges, capture and refund, and what the provider keeps.
- Worked designsscalability
Design search engine for the interview
Design search engine for the interview. Crawl, index, and query as three systems, plus typeahead as a smaller, hotter index.
Design web crawler for the interview
Design web crawler for the interview. The frontier, politeness, the seen-set, freshness versus coverage, and a Bloom filter.
- Consensus and coordinationconcurrency
Distributed locks for system design interviews
Distributed locks for system design interviews. Fencing tokens, expiry, the cases Redis misses, and the work that must not run twice.
- Consensus and coordinationconsistency
Distributed transactions for interviews
Distributed transactions for interviews. Why one BEGIN is gone, how outbox, saga, and 2PC differ, and a charge that must match a ledger row.
- Caching and storagesearch
Elasticsearch for system design interviews
Elasticsearch for system design interviews. The inverted index, relevance versus exact match, and why the search cluster sits beside the source of truth.
- Architecturearchitecture
Event driven architecture for interviews
Event driven architecture for interviews. Facts that already happened, consumers you can add later, and when a request-response call is clearer.
- Consensus and coordinationavailability
Fault tolerance for system design interviews
Fault tolerance for system design interviews. Name the fault first, then retries, isolation, bulkheads, and a store or queue that can lose a node.
- Caching and storagedatabases
Graph databases for system design interviews
Graph databases for system design interviews. Neighbour walks that explode in SQL, the write cost you take on, and when a recursive query is enough.
- Observability and opsscalability
Horizontal scaling for system design interviews
Horizontal scaling for system design interviews. When to add clones behind a balancer, when data refuses to clone, and when a bigger box is enough.
Interview study guide for system design
An interview study guide that stays one page: topics that show up in every loop, topics that can wait, and a daily review that keeps it honest.
- Consensus and coordinationconsistency
Leader election for system design interviews
Leader election for system design interviews. Why one writer exists, how leases and fencing stop a dead leader, and how to draw split brain.
- Caching and storagecaching
LRU cache for system design interviews
LRU cache for system design interviews. A hash map plus a doubly linked list, constant-time get and put, and the scan that flushes the working set.
- Architecturearchitecture
Microservices vs a monolith in interviews
Microservices vs a monolith in interviews. Independent deploys, the network cost, and when a modular monolith is the honest split.
- Observability and opsobservability
Observability for system design interviews
Observability for system design interviews. Logs, metrics, and traces as three signals, the question each answers, and cardinality that blows a bill.
Quorum reads and writes in system design
Quorum reads and writes in system design. The inequality that keeps a quorum safe, sloppy quorums, and what you give up for extra availability.
- Consensus and coordinationconsistency
Raft consensus for system design interviews
Raft consensus for system design interviews. Leader, log, and majority, an election after a crash, and what Raft still does not give you.
- Caching and storagecaching
Redis for system design interviews
Redis for system design interviews. Name the structure, name the job, and do not treat persistence as a database you can rebuild from.
- Traffic and reliabilitynetworking
Reverse proxy for system design interviews
Reverse proxy for system design interviews. TLS, buffering, caching, and the mistakes that turn the hop into a bottleneck.
- Architecturearchitecture
Sagas for distributed transactions, explained
Sagas for distributed transactions, explained. Choreography versus orchestration, compensating steps, and why a saga is not two-phase commit.
- Observability and opsobservability
SLO design for system design interviews
SLO design for system design interviews. How an SLI, an SLO, and an SLA differ, and how an error budget changes the week's work.
A software interview study plan by topic
A software interview study plan that splits hours between coding and design, for a two-week notice and for a three-month search.
- Architecturearchitecture
Stateless services for system design interviews
Stateless services for system design interviews. What the word excludes, where session state lives, and when a stateful service is the honest answer.
- Architecturearchitecture
Stream processing for system design interviews
Stream processing for system design interviews. Event time versus processing time, windows and watermarks, and when a join cannot wait for a batch.
System design interview: what to drill first
A system design interview scores coverage under a clock. What to clarify, when to estimate, and how the cards map onto the hour.
System design mock interview, run at home
Run a system design mock interview at home. Pick a prompt, record the hour, grade against key points, and turn gaps into cards.
System design preparation that fits in a month
System design preparation that fits a working month: a four-week calendar, what to drill first, and when to stop adding cards.
- Caching and storagedatabases
Time series databases for interviews
Time series databases for interviews. Append-heavy writes, range queries, rollups, and why a general store starts to hurt at ingest.
- Observability and opsobservability
Tracing in system design interviews
Tracing in system design interviews. A trace id that follows the request, spans and sampling, what traces miss, and OpenTelemetry as vocabulary.
- Consensus and coordinationconsistency
Two phase commit for system design interviews
Two phase commit for system design interviews. Prepare and commit, why the coordinator cannot vanish, and when a saga is the better story.
Back-of-the-envelope estimation for system design
A step-by-step method for back-of-the-envelope estimation in system design interviews, with worked QPS and storage examples and the numbers to memorise.
CAP theorem explained for interviews
CAP theorem explained for interviews. What the theorem actually says, why a partition forces a choice, and how to state it without the usual mistakes.
Consistent hashing explained for the interview
Consistent hashing explained for the interview. Why hash-mod-N remaps most keys, how the ring moves about 1/N, and what virtual nodes are for.
Database partitioning explained simply
Database partitioning explained simply. How to pick a shard key, what a hot key does, and how partitioning differs from a single replica.
Design news feed for the interview
Design news feed for the interview. Fanout on write versus fanout on read, what to cache, and how ranking stays out of the storage path.
Design URL shortener for the interview
Design URL shortener for the interview. The read path, the write path, how short codes are minted, and a capacity estimate you can redo.
FSRS vs SM-2: which scheduler should you study with?
FSRS vs SM-2 for flashcard study. What each does with a rating, what a lapse costs you, where SM-2 is still fine, and why MetaStack schedules with FSRS.
Idempotency keys in system design
Idempotency keys in system design. How a client retries a payment or a create without doubling the effect, and where the key has to be stored.
Load balancing for system design interviews
Load balancing for system design interviews. Round robin, least connections, and consistent hashing, and when layer 4 and layer 7 change the answer.
Message queues for system design interviews
Message queues for system design interviews. What a queue is for, how it differs from a log, and which delivery guarantee you should claim.
PACELC explained for system design
PACELC explained for system design. What the extra letter adds when the network is healthy, and how to choose latency or consistency on a whiteboard.
Rate limiting for system design interviews
Rate limiting for system design interviews. Token bucket, leaky bucket, and fixed windows, plus how a limit stays correct across more than one server.
Spaced repetition for system design: why it works
Why system design knowledge fades, why re-reading does not fix it, and how spaced repetition for system design turns ratings into a three-week schedule.
SQL vs NoSQL for system design interviews
SQL vs NoSQL for system design interviews. When a relational table is the right store, when a key-value or document store wins, and how to say the tradeoff.
System design interview flashcards that actually stick
How to make system design interview flashcards that train the answer you give out loud, not recognition. Card shapes, a grading rubric, and a daily routine.