<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Alex U</title>
    <description>The latest articles on DEV Community by Alex U (@alexu8007).</description>
    <link>https://dev.to/alexu8007</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3929741%2F185a0b92-5b38-423e-a5b4-fb8fdacf7bec.png</url>
      <title>DEV Community: Alex U</title>
      <link>https://dev.to/alexu8007</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/alexu8007"/>
    <language>en</language>
    <item>
      <title>Open-sourced ZenithDB, Fastest DB for Agent Traces in Rust</title>
      <dc:creator>Alex U</dc:creator>
      <pubDate>Wed, 13 May 2026 16:34:42 +0000</pubDate>
      <link>https://dev.to/alexu8007/open-sourced-zenithdb-fastest-db-for-agent-traces-in-rust-4ad6</link>
      <guid>https://dev.to/alexu8007/open-sourced-zenithdb-fastest-db-for-agent-traces-in-rust-4ad6</guid>
      <description>&lt;p&gt;Hey Guys!&lt;/p&gt;

&lt;p&gt;I just open-sourced &lt;a href="https://github.com/Polarityinc/zenith" rel="noopener noreferrer"&gt;Zenith&lt;/a&gt;, a columnar database engine purpose-built for AI agent traces.&lt;/p&gt;

&lt;h2&gt;
  
  
  why I built it
&lt;/h2&gt;

&lt;p&gt;Got mad with Clickhouse and couldn't find a OSS alternative for traces. tbh Agent traces are weird. they're long, sparse, high-cardinality JSON with bursty ingest, late annotations, and fat text fields. existing backends are built for short structured spans and pay a 10-100x cost on this workload.&lt;/p&gt;

&lt;p&gt;Built for things ppl acc do with Agent Traces:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;load a full trace tree&lt;/strong&gt; (B1): pull every span for a &lt;code&gt;trace_id&lt;/code&gt;, e.g., "show me everything that happened in this user's session". 571μs p95 at 1B rows&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;filter by attribute&lt;/strong&gt; (B2): "find spans where &lt;code&gt;model=claude-opus&lt;/code&gt; and &lt;code&gt;tokens &amp;gt; 4000&lt;/code&gt;". 4.3k μs p95 at 1B rows&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;full-text search over prompts and outputs&lt;/strong&gt; (B3): grep across millions of tool calls, prompts, and completions&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;JSONPath into nested attributes&lt;/strong&gt; (B6): query arbitrary JSON shapes without pre-defining a schema&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;group by model / tool / user&lt;/strong&gt; (B8): cost and token aggregations across a billion spans in 6.8ms&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  try it (30 seconds)
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;curl &lt;span class="nt"&gt;-fsSL&lt;/span&gt; https://raw.githubusercontent.com/Polarityinc/zenith/main/install.sh | sh
zen serve &lt;span class="nt"&gt;--config&lt;/span&gt; examples/zenithdb.dev.toml
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  links
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;repo: &lt;a href="https://github.com/Polarityinc/zenith" rel="noopener noreferrer"&gt;github.com/Polarityinc/zenith&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;Benchmarks &lt;a href="https://zenith.polarity.so" rel="noopener noreferrer"&gt;Zenith&lt;/a&gt;
&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;it's alpha. core engine is feature-complete but on-disk format and wire protocols may shift before 1.0. stars, issues, and PRs all welcome :)&lt;/p&gt;

</description>
      <category>database</category>
      <category>ai</category>
      <category>rust</category>
      <category>programming</category>
    </item>
  </channel>
</rss>
