<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
        <title>Ignacio Zenteno Smith — Blog</title>
        <link>https://izentenosmith.github.io/blog</link>
        <description>Technical write-ups on systems I've designed and run, and the trade-offs I'd revisit.</description>
        <lastBuildDate>Tue, 10 Nov 2026 15:50:44 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>https://github.com/jpmonette/feed</generator>
        <language>en</language>
        <image>
            <title>Ignacio Zenteno Smith — Blog</title>
            <url>https://izentenosmith.github.io/og.png</url>
            <link>https://izentenosmith.github.io/blog</link>
        </image>
        <copyright>© 2026 Ignacio Zenteno Smith</copyright>
        <atom:link href="https://izentenosmith.github.io/rss.xml" rel="self" type="application/rss+xml"/>
        <item>
            <title><![CDATA[Modeling configurable resolution workflows in PostgreSQL]]></title>
            <link>https://izentenosmith.github.io/blog/configurable-workflows-in-postgres</link>
            <guid isPermaLink="false">https://izentenosmith.github.io/blog/configurable-workflows-in-postgres</guid>
            <pubDate>Sat, 24 Oct 2026 12:00:00 GMT</pubDate>
            <description><![CDATA[How a ticketing platform lets admins define who fixes what, in which order and how fast, using a catalog/configuration/instance split and a closure table for the task graph.]]></description>
            <category>postgresql</category>
            <category>data-modeling</category>
            <category>workflows</category>
            <category>sql</category>
        </item>
        <item>
            <title><![CDATA[A monitoring stack as code on a two-node Docker Swarm]]></title>
            <link>https://izentenosmith.github.io/blog/monitoring-stack-as-code-on-docker-swarm</link>
            <guid isPermaLink="false">https://izentenosmith.github.io/blog/monitoring-stack-as-code-on-docker-swarm</guid>
            <pubDate>Sat, 24 Oct 2026 12:00:00 GMT</pubDate>
            <description><![CDATA[Grafana Alloy, Prometheus, Loki and Grafana watching hosts, containers, a JVM API, a Node frontend and three databases, with dashboards, alerts and DORA metrics all provisioned from git.]]></description>
            <category>observability</category>
            <category>grafana</category>
            <category>prometheus</category>
            <category>loki</category>
            <category>docker</category>
        </item>
        <item>
            <title><![CDATA[Blue/green deploys on a single VM with Docker Compose and Nginx]]></title>
            <link>https://izentenosmith.github.io/blog/blue-green-on-a-single-vm</link>
            <guid isPermaLink="false">https://izentenosmith.github.io/blog/blue-green-on-a-single-vm</guid>
            <pubDate>Sat, 24 Oct 2026 12:00:00 GMT</pubDate>
            <description><![CDATA[Zero-downtime releases and one-command rollbacks without Kubernetes, using two full Compose stacks, a symlinked Nginx upstream and a switch script that refuses to break the proxy.]]></description>
            <category>deployment</category>
            <category>nginx</category>
            <category>docker</category>
            <category>linux</category>
        </item>
        <item>
            <title><![CDATA[Designing a nightly ML scoring pipeline on Azure]]></title>
            <link>https://izentenosmith.github.io/blog/nightly-ml-scoring-on-azure</link>
            <guid isPermaLink="false">https://izentenosmith.github.io/blog/nightly-ml-scoring-on-azure</guid>
            <pubDate>Sat, 24 Oct 2026 12:00:00 GMT</pubDate>
            <description><![CDATA[A lakehouse, an MLOps loop and a fully private network for scoring and explaining around ten million financial transactions every night, plus the sizing details that are easy to get wrong.]]></description>
            <category>azure</category>
            <category>databricks</category>
            <category>mlops</category>
            <category>mlflow</category>
            <category>networking</category>
        </item>
        <item>
            <title><![CDATA[Inventory-driven network monitoring with Prometheus HTTP service discovery]]></title>
            <link>https://izentenosmith.github.io/blog/inventory-driven-network-monitoring</link>
            <guid isPermaLink="false">https://izentenosmith.github.io/blog/inventory-driven-network-monitoring</guid>
            <pubDate>Sat, 24 Oct 2026 12:00:00 GMT</pubDate>
            <description><![CDATA[A small Go service turns a device inventory into Prometheus targets, so every switch is pinged and polled over SNMP, and every metric carries its model, cabinet and map coordinates.]]></description>
            <category>prometheus</category>
            <category>grafana</category>
            <category>snmp</category>
            <category>go</category>
            <category>observability</category>
        </item>
    </channel>
</rss>