<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Systems Garden</title><description>Long-form systems engineering essays and lightweight notes on kernels, performance, compilers, and distributed systems.</description><link>https://draft-blog-captainirs-dev.pages.dev/</link><item><title>A tiny eBPF note: start with tracepoints before kprobes</title><link>https://draft-blog-captainirs-dev.pages.dev/notes/ebpf-first-tracepoint/</link><guid isPermaLink="true">https://draft-blog-captainirs-dev.pages.dev/notes/ebpf-first-tracepoint/</guid><description>If the goal is learning or getting a first signal quickly, tracepoints are often the cleaner entry point than kprobes. The interface is more stable, you spend less time chasing symbol details, and you still get a lot of practical observability value.</description><pubDate>Tue, 07 Apr 2026 00:00:00 GMT</pubDate></item><item><title>Why your ConfigMap changes are not showing up in the container yet</title><link>https://draft-blog-captainirs-dev.pages.dev/notes/configmap-reload-delay/</link><guid isPermaLink="true">https://draft-blog-captainirs-dev.pages.dev/notes/configmap-reload-delay/</guid><description>The short-form lane exists for useful operational fragments. If your pod mounts a ConfigMap as a volume, kubelet refresh is not an immediate push path. There is a sync period and cache behavior in play, so a change may take longer than you intuitively expect.</description><pubDate>Fri, 03 Apr 2026 00:00:00 GMT</pubDate></item><item><title>What Does It Take to Transfer 1 Byte Over Wi-Fi Without an OS?</title><link>https://draft-blog-captainirs-dev.pages.dev/posts/one-byte-over-wifi/</link><guid isPermaLink="true">https://draft-blog-captainirs-dev.pages.dev/posts/one-byte-over-wifi/</guid><description>A long-form post shape that can carry deep systems work, experiments, and the breadcrumbs needed to make the story discoverable later.</description><pubDate>Wed, 01 Apr 2026 00:00:00 GMT</pubDate></item></channel></rss>