05Design patternsPublish/subscribe messaging
05 · Reusable pattern
Publish/subscribe messaging
Design topic ownership, durable offsets, consumer groups, ordering, and replay.Lesson spine
What you need to understand.
Publish/subscribe decouples a producer from independent reactions while a durable log enables replay and scalable consumer groups.
Topics and partitions
A topic is split into ordered logs; the partition key defines which records keep order and which work can run in parallel.
Consumer groups
Assign each partition to one consumer within a group while separate groups receive their own copy of the event stream.
Offsets
Commit progress only after the side effect reaches its idempotent boundary; crash before commit means redelivery.
Rebalancing
Pause or fence old owners, transfer partitions, and resume from committed offsets without concurrent processing.
Fan-out
Email, analytics, fraud, and webhook groups can evolve independently from the producer.
CDC
Read source-database changes from its log to publish reliable projections without fragile application dual writes.
Before the boxes
Frame the decision.
What must work
Design topic ownership, durable offsets, consumer groups, ordering, and replay.
What changes the design
Throughput · record size · partitions · groups · retention · lag
What owns the truth
Identify the component that commits authoritative state, then separate synchronous confirmation from derived work.
What stays simple
Do not add global coordination, multi-region writes, or a specialized store until a requirement earns the complexity.
Architecture map
Trace ownership, not just traffic.
Follow the decision from left to right. Every arrow should have a reason.
Writes topic
Appends partition
Protects log
Assigns partitions
Processes offset
Rebalances ownership
Extends replay
Walk one representative request across every arrow. Say whether the handoff is synchronous or asynchronous, what identity makes a retry safe, and which step changes authoritative state.
- 01
Publisher — Writes topic Define the output contract before moving to the next owner.
- 02
Broker — Appends partition Define the output contract before moving to the next owner.
- 03
Replication — Protects log Define the output contract before moving to the next owner.
- 04
Consumer group — Assigns partitions Define the output contract before moving to the next owner.
- 05
Consumer — Processes offset Define the output contract before moving to the next owner.
- 06
Coordinator — Rebalances ownership Define the output contract before moving to the next owner.
- 07
Archive — Extends replay Confirm the result and emit the evidence needed to reconcile it.
Decision table
Make the trade-offs explicit.
| Decision | Defensible position | Cost to acknowledge |
|---|---|---|
| Primary mechanism | More partitions increase parallelism but weaken cross-key order and add coordination. | The stronger guarantee usually adds coordination, latency, state, or operational work. |
| Sync vs. async | Keep only correctness-critical confirmation synchronous. Move derived views, notifications, analytics, and cleanup behind a durable boundary. | Async work needs idempotency, lag monitoring, replay, and a product definition for partial completion. |
| Simple vs. scaled | Begin with one logical owner and a clear API. Partition or replicate only the resource proven to be the first bottleneck. | Migration requires stable identities, versioned contracts, backfill, and a rollback path. |
Failure review
Design the recovery path.
Topic-specific risk
Offset commit races during rebalance can duplicate or skip effects.
ResponsePersist enough identity and state to distinguish retry, resume, compensation, and operator repair.
Dependency timeout
A timeout is ambiguous: the remote side may have failed, succeeded, or still be running.
ResponseUse deadlines, bounded backoff with jitter, idempotency keys, and a status or reconciliation path.
Overload or skew
Average capacity can look healthy while a tenant, key, partition, region, or expensive request saturates one owner.
ResponseExpose queue depth and hot-key share, apply backpressure, isolate tenants, and degrade optional work before correctness.
Evidence + level bar
Prove the design can be operated.
Health of the promise
Measure user-visible latency or freshness, correctness drift, saturation, retry volume, and time to recover. Alert on the failed promise—not only CPU.
Complete and clear
Finish the happy path, identify the state owner, choose reasonable building blocks, and explain one scale mechanism.
Trade-offs and failure
Separate read and write paths, define consistency, explain partitioning, and make duplicate or partial failure safe.
Evolution and operations
Discuss multi-region boundaries, migration, tenant isolation, capacity, observability, and how the architecture changes over time.
Interview language
Open the deep dive with a claim.
“For Publish/subscribe messaging, the decision I want to make explicit is this: More partitions increase parallelism but weaken cross-key order and add coordination. I’ll trace the state-changing path first, show where the result becomes durable, then test the design against the highest-risk failure and our target scale.”
08 · Retrieval check
Can you defend it without the page?
- For Publish/subscribe messaging, where is the correctness boundary and which failure would you test first?
- Which component owns committed truth, and what event or response proves the commit?
- Where is the first scaling or coordination bottleneck under the stated envelope?
- What happens after an ambiguous timeout or duplicate operation?
- Which complexity would you remove at one hundredth of the scale?