<?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: nexora</title>
    <description>The latest articles on DEV Community by nexora (@nexoradev).</description>
    <link>https://dev.to/nexoradev</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%2F3978407%2Fb48a3220-3433-40d7-9c3e-c3f018c49b4d.png</url>
      <title>DEV Community: nexora</title>
      <link>https://dev.to/nexoradev</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/nexoradev"/>
    <language>en</language>
    <item>
      <title>I built a rate limiter that survives process restarts — no Redis required</title>
      <dc:creator>nexora</dc:creator>
      <pubDate>Thu, 11 Jun 2026 01:41:28 +0000</pubDate>
      <link>https://dev.to/nexoradev/i-built-a-rate-limiter-that-survives-process-restarts-no-redis-required-g4n</link>
      <guid>https://dev.to/nexoradev/i-built-a-rate-limiter-that-survives-process-restarts-no-redis-required-g4n</guid>
      <description>&lt;p&gt;Every time I needed rate limiting in a Python service, the answer was the same: add Redis. Just to count requests.&lt;br&gt;
So I built Flint. Embedded, persistent, zero infrastructure.&lt;br&gt;
pythonimport flint&lt;/p&gt;

&lt;p&gt;limiter = flint.Limiter(data_dir=".flint")&lt;br&gt;
limiter.limit("api:user-42", rate=100, per="1m")&lt;/p&gt;

&lt;p&gt;if limiter.allow("api:user-42"):&lt;br&gt;
    process_request()&lt;br&gt;
Counters persist in .flint/ and survive restarts. Something no in-memory solution does.&lt;/p&gt;

&lt;p&gt;What makes it different&lt;br&gt;
SlowAPI resets on restart. Redis requires infrastructure. nginx only works at HTTP layer.&lt;br&gt;
Flint is persistent, embedded, and works anywhere — including AI APIs with cost-based limiting:&lt;br&gt;
pythonresult = limiter.check("ai:user-42", cost=250)&lt;br&gt;
FastAPI middleware, Prometheus export, millisecond precision, and atomic multi-limit checks included.&lt;br&gt;
Core is Rust via PyO3. Token bucket check: ~560µs.&lt;/p&gt;

&lt;p&gt;bashpip install flint-limiter&lt;/p&gt;

&lt;p&gt;GitHub: &lt;a href="https://github.com/BuildByNexora/Flint" rel="noopener noreferrer"&gt;https://github.com/BuildByNexora/Flint&lt;/a&gt;&lt;/p&gt;

</description>
      <category>programming</category>
      <category>rust</category>
      <category>ratelimiting</category>
      <category>opensource</category>
    </item>
    <item>
      <title>I built an embedded scheduler in Rust because I was tired of adding Redis just to run a background job</title>
      <dc:creator>nexora</dc:creator>
      <pubDate>Wed, 10 Jun 2026 21:53:16 +0000</pubDate>
      <link>https://dev.to/nexoradev/i-built-an-embedded-scheduler-in-rust-because-i-was-tired-of-adding-redis-just-to-run-a-background-597b</link>
      <guid>https://dev.to/nexoradev/i-built-an-embedded-scheduler-in-rust-because-i-was-tired-of-adding-redis-just-to-run-a-background-597b</guid>
      <description>&lt;p&gt;Every time I needed to run a background job in a Python app, the answer was always the same: add Redis, add Celery, add a worker process, add a broker.&lt;br&gt;
Just to send an email every morning.&lt;br&gt;
So I built Kron. It runs embedded inside your Python process — no Redis, no broker, no daemon. Timers persist across restarts. If the process crashes, Kron recovers everything from a local append-only log.&lt;br&gt;
pythonimport kron&lt;/p&gt;

&lt;p&gt;kron.schedule("email_digest", cron="0 8 * * *", fn=send_digest)&lt;br&gt;
kron.schedule("cleanup", every="30m", fn=cleanup_temp_files)&lt;/p&gt;

&lt;p&gt;kron.start(data_dir=".kron")&lt;br&gt;
That's it. No infrastructure.&lt;br&gt;
You can inspect everything from the CLI:&lt;br&gt;
bashkron job list&lt;br&gt;
kron job status email_digest&lt;br&gt;
kron job history email_digest&lt;br&gt;
The core is Rust. Python bindings via PyO3. It also has a distributed server mode backed by OpenRaft with worker leases, fencing tokens, and a tamper-evident audit log with SHA256 hash chaining.&lt;br&gt;
pip install kron-scheduler&lt;br&gt;
GitHub: &lt;a href="https://github.com/BuildByNexora/Kron" rel="noopener noreferrer"&gt;https://github.com/BuildByNexora/Kron&lt;/a&gt;&lt;br&gt;
Still alpha — but already useful for small services, local agents, and edge devices.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>python</category>
      <category>opensource</category>
      <category>rust</category>
    </item>
  </channel>
</rss>
