Arcturus documentation

Arcturus is reliability and management software for individual machines, home clusters, fleets and data centers. It builds on the Rubian programmable operating environment and the Kestowv systems foundation. Its scope includes predictive hardware maintenance at the machine level. Arcturus extends predictive maintenance analytics across the data-center stack, including the power, cooling, networking and facility systems on which the machines depend.

Product scope and current documentation#

Rubian provides the programmable operating environment, Kestowv the systems foundation, and Valen the AI. Core is the user- and agent-facing application environment; selected interface work is intended to contribute to the broader operating-system research. The Valux project explores the broader operating system across personal and enterprise computing.

Arcturus applies that foundation to reliability and management across machines and fleets: organized discovery and work, reusable operations, and coordinated administration across the operating environment. Current component behavior includes Rubian's startup inventory, project records and agent collaboration. Reducing unnecessary system work is a product objective; deployment-specific resource and performance results require their own evaluation.

The intended predictive-maintenance scope covers utility power through UPS and PDU, cooling equipment, switches and optics, servers, storage, GPUs, and fire, leak and access systems. The purpose of bringing those observations together is to support decisions about the reliability of the whole operating environment. The scope guide describes the planned coverage and its relationship to the current demonstrations.

Arcturus is under active development. The public demonstrations currently cover data-center observations, machine inspection and operational work. These are reliability and management workflows; the demonstrations show a narrower set of workflows than the full product scope. The guides below explain those available interfaces. Installation, hardware support and enterprise administration require documentation tied to the supplied Arcturus release.

Choose an investigation#

The Data center demonstration begins with the facility: cooling, power, networking, racks and the services placed within them. Use it to investigate a condition whose consequences extend beyond one machine. The data-center guide follows a cooling observation through dependency review and a maintenance plan.

The Machines view begins with a selected host. Use it to examine inventory, current counters, available sensor observations, findings and related incident work. The machine guide explains how to retain the observation associated with an incident.

Explore the public interfaces#

Open Arcturus for data centers or the machine-level preview. These demonstrations use representative and synthetic data. No live equipment is connected. Acknowledgements, maintenance plans and simulated incidents belong to the demonstration in your browser.

The demonstration lets you evaluate the investigation workflow: which source produced a reading, when it was observed, what depends on the affected equipment, and what the operator plans to do next. It does not produce a trained failure forecast or operate customer equipment.

Read a condition in context#

A reading needs a unit, timestamp and source to be useful. A threshold crossing can justify investigation without establishing a failed component. An unavailable observation requires follow-up rather than a healthy label. When a concern spans several assets, examine the dependency relationship and the service impact separately.

Use Observations and troubleshooting to interpret stale data, missing sensors, local state and export behavior.

Plan an evaluation on real equipment#

A connected evaluation needs an agreed inventory, supported observation sources, access boundaries, alert criteria and operating responsibilities. Specify which work is observational and which actions require an operator. Include the expected behavior when a source stops reporting and the evidence needed to close an incident.

Contact Valen Systems with the data-center scope and the operational question you need to answer. Connected deployment instructions depend on the supplied environment and its approved integrations.