Blog

The PrexorCloud blog covers release notes, architectural decisions we think are worth a long-form explanation, and engineering deep-dives into the parts of the system we are most often asked about. Posts are written by the people building PrexorCloud and live in the same git repository as the source — issue a PR if you want to suggest a correction.

If you are looking for a structured changelog, see the changelog instead. The blog is for stories the changelog is too terse to tell.

Launch posts

Three posts ship alongside the v1.0 launch. They are listed below in the recommended reading order — start with the announcement, then the vision piece if you want to know why this project exists, then the technical deep-dive if you want to know how one of its more interesting layers works.

PrexorCloud v1.0 — A cloud system built for Minecraft

The launch post. What v1.0 is, who it is for, what it includes, and what it deliberately leaves out. If you are evaluating PrexorCloud against CloudNet 4 or SimpleCloud V2, this is the right starting point — it covers the headline features, the Apache 2.0 licence, the release artifacts, the cosign-signed install path, and the five things v1 deliberately leaves out of scope.

Why we built another cloud orchestrator

A longer essay on the framing question every Minecraft operator asks when they meet a new orchestrator: did we really need another one? The post walks through six trade-offs PrexorCloud picked deliberately — single-tenant by design, MongoDB plus Valkey as the data layer, signing as a v1 default rather than a v2 opt-in, active-active HA without leader election, no per-instance containers, no SSO — and argues why those choices add up to a different product category from the existing options.

Daemon-side modules — how Layer 7 unlocks node-local extension

A walk through the design of daemon-side modules: the new DaemonModule interface, the controller-side ModuleDistributor that fans installs out to every connected daemon over gRPC, the event bridge that lets a module on a host subscribe to controller events without polling, and the instance-lifecycle hooks (onInstanceStarting / Started / Stopping / Stopped) that make node-local logic — JVM args, env mutation, profiling injection, custom crash handling — possible without forking the daemon. The canonical reference is the daemon module concept page; the post puts the design choices in context, with sequence and state diagrams of the install and lifecycle flows.

Subscribing

There is no email list yet. The blog ships as part of this site, so any RSS reader pointed at /blog/rss.xml will pick up new posts as they land. GitHub watchers of the main repository get release notifications, which line up with the bigger blog posts.