update project entries

2026-07-08 10:26:45 -07:00
1 parent 7c1c049148
commit 051f95f00d
@@ -0,0 +1,45 @@
+---
+title: Reforger Mods API
+slug: reforger-mods-api
+summary: A public, unofficial REST API for Arma Reforger Workshop metadata, built with caching, background refresh jobs, rate limiting, and production reliability in mind.
+label: Project
+date: 2026-07-06
+status: Active
+featured: true
+order: 1
+github: https://github.com/SowinskiBraeden/ReforgerWorkshopAPI
+live: https://api.reforgermods.net
+tags: Go, REST API, Web Scraping, Observability, Documentation
+---
+
+## Overview
+
+The Arma Reforger Workshop is primarily a website, not a developer API. Panels, hosting tools, bots, and dashboards that need structured mod data have no clean way to get it. I built Reforger Mods API to fill that gap.
+
+It is a public REST API, live at [api.reforgermods.net](https://api.reforgermods.net), that fetches and normalizes Workshop metadata into predictable, documented responses for developers. After the most recent relaunch, it started receiving identified traffic from third-party, which makes it one of the first projects I have built that is genuinely used beyond my own stack.
+
+## What It Does
+
+- Search and list Workshop mods with structured, filterable responses
+- Retrieve detailed mod metadata normalized from raw Workshop pages
+- Cache responses to avoid hammering upstream on repeated lookups
+- Serve stale data while background refresh jobs revalidate it
+- Return `202 Accepted` on cold cache misses so consumers are not blocked waiting on a full upstream scrape
+
+## Reliability and API Design
+
+A scraper-backed API has an obvious weak point: if the upstream source is slow or flaky, every request suffers. I treated reliability as a first-class concern from the start.
+
+Cached responses return fast. When data is stale, it gets served immediately while a refresh job runs in the background. On a cold miss, the API returns `202 Accepted` with a job reference instead of making the caller wait. Rate limiting and defensive scraper error handling mean that one bad upstream page does not cascade into a broken API response.
+
+Internal metrics track traffic, cache hit rates, refresh job behavior, and errors, so I can actually tell what the service is doing in production rather than guessing.
+
+## Documentation
+
+The API includes public-facing documentation covering all endpoints, response shapes, cache behavior, and how to handle `202 Accepted` responses correctly. Getting the docs right felt important, if external tools are depending on this, they need to be able to understand it without digging through source code.
+
+Live API and docs: [api.reforgermods.net](https://api.reforgermods.net)
+
+## Notes
+
+This project sits at the intersection of a few things I am consistently interested in: backend infrastructure, reliability engineering, and practical tooling for game communities. It is also the first project where I have had to think seriously about backwards compatibility and defensive API design — once external consumers exist, breaking changes have real consequences. That constraint pushed me to build it more carefully than I would have if it were purely internal.