What's New in Rust's "Devblog 185 - Post Mortem" Update

Two days after Devblog 185 shipped in November 2017, Facepunch posted a short and unusually candid post-mortem with a one-line summary: "Why we rolled back." For the second week running, a patch-day bug had bitten hard — this time in the brand-new Workbench Experimentation system — and Helk used the occasion to question the studio's whole weekly-patch cadence.

What Broke

Helk's own workbench code contained a nasty timing bug: if the server initiated a save while any player was mid-experiment, the save was prevented from completing. The team spent all night trying to rescue affected servers — going as far as writing injector programs to modify the server's memory and bring its save functionality back to life — but nothing worked. Any server that dodged the unlucky timing was fine; servers that happened to save during that roughly five-second window were, in Helk's words, screwed, and had to be rolled back.

Second-Guessing Weekly Patches

The more interesting half of the post was the reflection. Helk admitted that incidents like this made him second-guess the weekly patch approach entirely. Despite a good track record, he argued this kind of failure wouldn't be acceptable as Rust moved out of early access — and that the weekly rhythm sometimes forced content into patches before it was ready for primetime, giving players sub-standard work. The proposed fix: re-evaluate the process and consider a feature-locked patch on a separate branch, tested for a solid week before being pushed live. The post closed with a simple "We're so sorry."

In hindsight this post-mortem marked a real turning point — the thinking here fed directly into Rust's eventual move away from weekly content patches toward the tested monthly-update cycle players know today.