Knowledge section

Patterns

Reusable implementation patterns, failure boundaries, and architecture moves for production systems.

intermediate

Bulkhead Isolation

Isolate resources so one failing dependency, tenant, or workload cannot consume capacity needed by the rest of the system.

reliabilityresilienceresource-isolation

foundation

Cache-Aside

A caching pattern where the application reads from cache first, falls back to the database on misses, and then populates the cache.

cachingperformancescalability

intermediate

Circuit Breaker

Stop calling an unhealthy dependency for a period of time so failures do not cascade through the system.

reliabilityresilienceservice-communication

intermediate

Connection Registry

Track which realtime gateway currently owns each connected user device so delivery workers can route live events without making gateways durable truth.

realtime-systemsmessagingrouting

intermediate

CQRS Read Model

Separate write models from read-optimized projections so systems can preserve source-of-truth writes while serving efficient queries.

cqrsread-modelsdata-modeling

intermediate

Cursor-Based Sync

Let clients recover missed updates by asking for durable records after their last known position instead of trusting live delivery alone.

messagingsynchronizationreliability

foundation

Dead-Letter Queue

Isolate messages that repeatedly fail processing so they do not block healthy event processing forever.

queuesreliabilityoperations

intermediate

Idempotent Consumer

Safely process duplicate messages by making consumer side effects converge to the same final result.

reliabilityevent-driven-architecturemessaging

advanced

Large Group Fan-Out Isolation

Keep massive group delivery workloads from starving normal messaging by separating queues, worker pools, limits, and backpressure policies.

messagingscalabilityfan-out

intermediate

Lease-Based Assignment

Temporarily reserve a scarce resource for one workflow so concurrent workers cannot assign it twice while the final confirmation is still pending.

distributed-systemsconcurrencyreliability

intermediate

Read Cursor Receipts

Represent read state as a user's latest read sequence in a conversation instead of writing one read receipt per message.

messagingread-modelsscalability

intermediate

Reconciliation Job

Periodically compare source-of-truth data with derived state and repair drift caused by missed, duplicated, or incorrectly processed updates.

reliabilitydata-consistencyrepair

foundation

Retry With Exponential Backoff And Jitter

Retry transient failures with increasing delays and randomness so recovery does not create a synchronized traffic spike.

reliabilityretriesdistributed-systems

advanced

Saga Pattern

Coordinate a multi-step distributed workflow using local transactions and compensating actions instead of one large distributed transaction.

distributed-transactionsworkflowreliability

intermediate

Sharded Counter

Split one logical counter into multiple partial counters to distribute write load for hot aggregates.

countersscalabilityhot-keys

intermediate

Transactional Outbox

Reliably publish events that correspond to committed database changes by storing publish intent inside the same local database transaction.

reliabilityevent-driven-architecturedata-consistency

intermediate

Upload-Then-Reference Media

Upload large media outside the message send path, then send a lightweight message that references the stored media object.

messagingstoragemedia