<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Http on Lorbic</title><link>http://localhost:1313/tags/http/</link><description>Recent content in Http on Lorbic</description><generator>Hugo</generator><language>en</language><lastBuildDate>Sun, 26 Jul 2026 19:00:00 +0530</lastBuildDate><atom:link href="http://localhost:1313/tags/http/index.xml" rel="self" type="application/rss+xml"/><item><title>API Design for Backend Systems</title><link>http://localhost:1313/api-design-for-backend-systems/</link><pubDate>Sun, 26 Jul 2026 19:00:00 +0530</pubDate><guid>http://localhost:1313/api-design-for-backend-systems/</guid><description>&lt;p>A few weeks ago I designed &lt;a href="http://localhost:1313/how-multi-tenant-saas-works/">Relay&lt;/a>, a multi-tenant AI API gateway, as a system design exercise. That post covered the database, the auth, the billing. It did not cover the thing every one of those systems is sitting behind: the API itself.&lt;/p>
&lt;p>Here is the question that post left open. When someone builds &lt;code>GET /v1/requests&lt;/code> to list a tenant&amp;rsquo;s API call history, what should that endpoint actually look like? Offset or cursor pagination? What does the response envelope contain? What happens when Relay needs to add a filter that does not fit in a query string? What does a webhook payload look like when Relay tells a tenant &amp;ldquo;your batch job finished&amp;rdquo;? None of that is specific to AI gateways. It is the same decisions every backend API makes, made badly often enough that &amp;ldquo;REST API&amp;rdquo; has become a phrase that means &amp;ldquo;JSON over HTTP, structure unspecified.&amp;rdquo;&lt;/p></description></item></channel></rss>