Dead Letter Queue in Microservices: Handling Poison Messages
Learn how to design a dead letter queue in microservices: classify poison messages, bound message retry, enrich failures, and run safe DLQ replay with owners.
13 min read
Archive
All articles
295 published
Debezium PostgreSQL CDC: Slots, Snapshots and Safe Consumers
Set up Debezium PostgreSQL CDC on Kafka Connect: logical replication slots, snapshots, event envelopes, idempotent consumers and WAL-safe operations.
13 min read
Backward Compatible API Changes for Independently Deployed Services
A checklist for backward compatible API changes: safe vs breaking changes, tolerant readers, expand-and-contract, enums, Protobuf and event rules, and CI checks.
18 min read
gRPC Load Balancing in Kubernetes: Why Connections Stick to One Pod
Why Kubernetes Services pin gRPC's long-lived HTTP/2 connections to one pod, how to confirm it, and how to fix it with client-side or L7 proxy balancing.
12 min read
Rate Limiting a Monolith Without Redis: In-Memory vs Postgres
Compare in-memory and Postgres rate limiters for a monolith: algorithms, atomic SQL, process-count pitfalls, failure modes, and how to choose by topology.
14 min read
PgBouncer Transaction Pooling: Prepared Statements and Session State
How PgBouncer transaction pooling works, why prepared statements and session state break, how to audit your app, and how to size and monitor pools.
13 min read
Outbox Relay Lag: Measuring and Tuning Each Stage of Event Delay
Break outbox relay lag into commit, capture, publish and consume segments, measure each one, monitor CDC pipelines, and tune the stage that breaks your budget.
17 min read
Nx Affected Database Migrations: Run Only Changed Schemas
Model each schema as an Nx project with a non-cached migrate target, pick safe base SHAs, and keep a full-run safety net so no environment drifts.
13 min read
Nginx Upstream Keepalive and Timeouts for Node.js Backends
Fix intermittent 502s between Nginx and Node.js by aligning nginx upstream keepalive with Node's keepAliveTimeout and scoping proxy timeouts per location.
16 min read
NestJS Graceful Shutdown: Draining Requests and Queues on SIGTERM
Go beyond enableShutdownHooks(): order readiness, HTTP connection draining, queue consumers and timeouts so NestJS pods exit cleanly on Kubernetes SIGTERM.
14 min read
Inbox Pattern in Microservices: Deduplicating Incoming Events
Build an inbox pattern for idempotent consumers: choose dedup keys, design a processed messages table, handle concurrency, retention, testing and trade-offs.
13 min read
FastAPI Dependency Injection: Scopes, Overrides and Testing
How FastAPI dependency injection resolves a per-request graph, manages app-lifetime resources, cleans up with yield, and stays testable with dependency_overrides.
14 min read