DEV BAK - TECH BLOG

Orchestrating page transitions in MPAs with the View Transitions API using CSS alone

Using the View Transitions API's cross-document mode, you can implement cross-fade and shared element transitions between pages on MPA and SSG sites with just two lines of CSS declarations and no JavaScript, while unsupported browsers automatically fall back to the default behavior, preserving the principle of Progressive Enhancement.

September 18, 202615 min read

DevOps

Three gaps that appear when moving from Helm to Kustomize — and how to fill them again: template functions, hooks, and tests

This guide specifically explains how to fill the gaps left by disappearing template functions, release hooks, and test capabilities when migrating from Helm to Kustomize, using tool and pattern combinations that are practically applicable in 2026.

September 13, 202621 min read

Backend

Keeping validation and documentation in sync from a single schema with Hono + Zod OpenAPI

This covers how to simultaneously generate TypeScript type inference, runtime validation, and OpenAPI documentation from a single Zod schema source using Hono's `@hono/zod-openapi`, as well as patterns for structurally reducing Schema Drift.

September 12, 202618 min read

frontend

Declaring Tooltip, Dropdown, and Popover Positions with CSS Alone Using CSS Anchor Positioning

Based on the core API and practical usage of CSS Anchor Positioning, this covers how to implement tooltips, dropdowns, and popovers declaratively with CSS alone—no JavaScript required—along with situations where a library is still necessary.

August 30, 202618 min read

DevOps

Argo Rollouts Canary Deployment: Letting Prometheus Metrics Drive Traffic Percentage Instead

Explains how to use Argo Rollouts' Canary strategy and AnalysisTemplate to automatically advance or roll back traffic steps based on Prometheus metrics, complete with real YAML examples, while also covering the limitations of basic Deployment rolling updates and common configuration pitfalls.

August 29, 202618 min read

Claude

Things I learned from implementing a Claude MCP server from scratch — tool definitions, handler routing, and schema validation

When implementing an MCP server, this explains step by step how tool definitions (name, description, JSON Schema), tools/list and tools/call handler routing, and Zod/Pydantic-based schema validation interlock within actual code flow. It also covers the criteria for choosing between stdio and Streamable HTTP transports.

August 28, 202618 min read

OpenSource

What happens when you port ESLint rules to Rust — Oxc AST, the visitor pattern, and the reality of JS plugins

From a practitioner's perspective, this article compares and analyzes the differences in AST structure, visitor pattern behavior, and the practical scope of the JS plugin alpha stage when porting ESLint rules to Rust in Oxlint.

August 26, 202619 min read

Architecture

Choreography vs. Orchestration — The difference you only see after implementing distributed transactions yourself

A step-by-step comparative analysis of how the Saga pattern works when implemented via Choreography versus Orchestration, including code-level differences and practical selection criteria, starting from the limitations of 2PC.

August 23, 202622 min read

Claude

Tying it all together in a single loop with Claude Tool Use — covering stop_reason branching, parallel calls, and turning errors into data

We explain the agentic loop of Claude Tool Use with a focus on stop_reason branching, parallel tool calls, and error datafication patterns, and cover production-level implementation methods using a competitor price monitoring pipeline as an example.

August 19, 202622 min read

Architecture

Quantifying the tradeoffs of tenant isolation across three layers: schema, database, and process

Compare the three isolation layers — schema, database, and process — in a multi-tenant SaaS architecture using code and metrics, and quantify the trade-offs so you can select a practically applicable combination based on cost, regulatory requirements, and operational overhead.

August 16, 202622 min read

Backend

Configuring fanout and retry in a single file with Cloudflare Workers + D1 + Queues

A practical guide with code showing how to co-locate an HTTP producer and queue consumer in a single Cloudflare Workers file, track job state with D1, and implement fanout, retries, and DLQ.

August 8, 202622 min read

Backend

A record of wiring up real-time broadcast without an external broker, using PostgreSQL LISTEN/NOTIFY and Bun WebSocket

How to implement real-time broadcasting without Redis by connecting PostgreSQL's built-in LISTEN/NOTIFY mechanism with Bun WebSocket, along with practical limitations discovered in production — including transaction commit guarantees and reconnection handling — explained with code.

August 7, 202617 min read