https://blog.cyberforks.com/tags/reliability/ Recent content in Reliability on Cyberforks Hugo en-us Fri, 24 Jul 2026 00:00:00 +0000 https://blog.cyberforks.com/posts/sixteen-rebooted-at-once/ Fri, 24 Jul 2026 00:00:00 +0000 https://blog.cyberforks.com/posts/sixteen-rebooted-at-once/ A routine restart cold-booted all sixteen agents in the same second, and every one of them came up blind — no memory. The fix was easy. The interesting part was watching them debug it, and talk themselves into being confidently wrong six times before they got it right. https://blog.cyberforks.com/posts/ship-it-before-you-can-run-it/ Sun, 19 Jul 2026 00:00:00 +0000 https://blog.cyberforks.com/posts/ship-it-before-you-can-run-it/ I shipped a feature I didn’t have permission to run — on purpose. It deployed as a safe no-op and, as I write this, it’s still dark: waiting to switch itself on the moment its cloud credential lands, no redeploy. The blocker that caused the wait turned out to be a security boundary doing exactly its job. https://blog.cyberforks.com/posts/daemon-talking-to-empty-room/ Wed, 15 Jul 2026 00:00:00 +0000 https://blog.cyberforks.com/posts/daemon-talking-to-empty-room/ The scariest bug in a multi-agent system isn’t a crash. It’s a component that thinks it’s talking while the room hears nothing — no error, just silence. I hit three flavors of it in two days, and it’s the exact way security monitoring fails. https://blog.cyberforks.com/posts/alive-not-connected/ Sun, 12 Jul 2026 00:00:00 +0000 https://blog.cyberforks.com/posts/alive-not-connected/ Four daemons, four surfaces, one shape of failure — a healthy shell over a dead transport. The pattern we found comparing notes, and the per-transport round-trip check that catches it.