Originally created by: fu351
Today each host hook pays engine import/startup cost per call. A warm, per-device background daemon that hosts the decision engine would remove that cost and unlock ambient monitoring — and it's the open foundation other collectors plug into.
What to build
Collector / ActivitySource entry-point interface + a local ambient event bus (mirrors the existing Rule/Detector/AuthProvider/AuditSink plugin pattern).Where to start: this needs a short design pass first (propose the IPC boundary, the daemon lifecycle, and the Collector interface) — it isn't slice-detailed yet, so comment with a design sketch before coding. Reuse the existing engine/registry; keep the daemon strictly off the hot decision path's correctness guarantees (it's a performance + observability layer).
Difficulty: hard / architectural — an epic; great for someone who wants ownership of a subsystem.
Building this? A few project invariants any change must respect:
New here? See CONTRIBUTING.md (github.com), and come chat in the Discord — happy to help you scope any of this.
Originally posted by: fu351
Closing as decomposed. This level-10 umbrella was too big to hand anyone, so it was split into buildable slices: [#236] (the activity bus), [#237] (the warm observe-only daemon itself), and [#238] (basic collectors). The design detail lives with those issues. If you want in on F-Monitor, start with [#236]; the daemon in [#237] builds on it.
Related
Tickets: #236
Tickets: #237
Tickets: #238
Ticket changed by: fu351