<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Anomaly-Detection on Pi Stack</title>
    <link>https://www.pistack.xyz/tags/anomaly-detection/</link>
    <description>Recent content in Anomaly-Detection on Pi Stack</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Mon, 15 Jun 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://www.pistack.xyz/tags/anomaly-detection/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Self-Hosted Log Anomaly Detection: Drain3 vs Loglizer vs LogDeep</title>
      <link>https://www.pistack.xyz/posts/2026-06-15-self-hosted-log-anomaly-detection-drain3-loglizer-logdeep/</link>
      <pubDate>Mon, 15 Jun 2026 00:00:00 +0000</pubDate>
      <guid>https://www.pistack.xyz/posts/2026-06-15-self-hosted-log-anomaly-detection-drain3-loglizer-logdeep/</guid>
      <description>&lt;h2 id=&#34;introduction&#34;&gt;Introduction&lt;/h2&gt;&#xA;&lt;p&gt;Modern applications generate terabytes of log data daily — HTTP access logs, application error logs, database query logs, system audit trails, and security event logs. Within this ocean of mostly routine entries, a handful of anomalous log lines may indicate a critical failure, a security breach, or a performance regression. Finding these needles in the log haystack is the core challenge of log anomaly detection.&lt;/p&gt;</description>
    </item>
    <item>
      <title>Self-Hosted Metric Anomaly Detection: Luminol vs EGADS vs Surus</title>
      <link>https://www.pistack.xyz/posts/2026-06-15-self-hosted-metric-anomaly-detection-luminol-egads-surus/</link>
      <pubDate>Mon, 15 Jun 2026 00:00:00 +0000</pubDate>
      <guid>https://www.pistack.xyz/posts/2026-06-15-self-hosted-metric-anomaly-detection-luminol-egads-surus/</guid>
      <description>&lt;h2 id=&#34;introduction&#34;&gt;Introduction&lt;/h2&gt;&#xA;&lt;p&gt;Modern infrastructure generates millions of metrics every minute — CPU utilization, request latency, error rates, queue depths, and hundreds more. Manually setting static alert thresholds (&amp;ldquo;alert if CPU &amp;gt; 80%&amp;rdquo;) breaks down at scale because what&amp;rsquo;s normal at 2 PM on Tuesday differs dramatically from normal at 3 AM on Sunday.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
