Skip to content

Analytics

The first thing a fleet needs is to be understood: what Dusk is running across your devices, how those workloads are behaving, and how that changes over time. Analytics is the payoff most people reach for first after embedding Dusk.

TODO - in progress. Fleet-wide analytics tooling is still being built. This page describes where it's headed; today you collect the raw signals per node with the diagnosis tools and aggregate them yourself.

Where it's headed

  • A live view of every node - its processes, versions, and health - across the whole fleet.
  • Metrics and structured logs streamed off each node and rolled up centrally.
  • Trends over time, so you can watch a single device - or a whole rollout - drift before it becomes an incident.

What you can do today

Each node already exposes the raw material: process listings (ps), program metadata, and structured log records. Pull them per node through the client interfaces and feed them into your own store while the managed pipeline matures.