Blog

Notes on building things that work.

Engineering, cloud, AI and product thinking — written for people who make decisions about technology.

Abstract illustration of Python asyncio task groups coordinating concurrent tasks with structured lifetimes and cancellation

Engineering31 Oct 2026· 10 min

Python asyncio TaskGroups: Structured Concurrency That Survives Cancellation

A production-focused guide to Python 3.11+ asyncio TaskGroups: child-task ownership, sibling cancellation, ExceptionGroup handling, deadlines, and safe fan-out patterns.

Read article

PostgreSQL primary and standby synchronizing logical replication slots to keep Debezium and other CDC consumers running after failover

Engineering31 Oct 2026· 11 min

PostgreSQL Logical Replication Failover: Keep CDC Running After Promotion

A production guide to PostgreSQL 17+ logical failover slots: configure slot synchronization, protect WAL, verify standby readiness, and recover CDC consumers safely after promotion.

Read article

Reliable Python backend architecture with API services, observability, and protected infrastructure

Engineering31 Oct 2026· 8 min

Building Reliable FastAPI Services in Production: Timeouts, Retries, Idempotency, and Observability

A production-focused guide to making FastAPI services resilient with deadlines, bounded retries, idempotency, database safeguards, backpressure, and observability.

Read article

Transactional outbox architecture connecting PostgreSQL, an event relay, Kafka, and downstream microservices

Engineering31 Oct 2026· 10 min

Reliable Event-Driven Microservices with the Transactional Outbox Pattern

A production-focused guide to the transactional outbox pattern: atomic PostgreSQL writes, Debezium CDC, Kafka delivery, idempotent consumers, and operational failure handling.

Read article