Skip to main content

System

Purpose​

System — the physical host's hardware topology (CPU sockets/cores/threads, NUMA nodes, hugepages) plus topology-aware optimization recommendations for individual VMs. This is distinct from System Health, which tracks live utilization rather than hardware layout.

Four tabs: CPU Topology, NUMA Topology, Memory & Hugepages, Optimization (badged with pending recommendation count).

When to use it​

  • To see the host's CPU socket/core/thread layout and NUMA node boundaries before pinning a VM's vCPUs
  • To check how many hugepages (2MB/1GB) are allocated and free, or allocate more for a memory-intensive VM
  • To review and apply topology-aware optimization recommendations (e.g. NUMA/CPU pinning) for a specific running VM
  • Prefer this page when the job matches the purpose above
  • Before placing latency-sensitive guests, to understand NUMA distances

How to get there​

  • Route / id: /app/system
  • Nav: Infrastructure → System (sidebar, command palette, or desktop nav)

Operate from the console (UX)​

Stat cards: total CPUs (sockets × cores), NUMA node count, total/available memory, 2MB hugepage total/free.

  1. CPU Topology — CPUs by socket, online/offline; hover for core, thread, NUMA node.
  2. NUMA Topology — per-node CPUs, memory total/free + bar, hugepage counts, inter-node distance matrix.
  3. Memory & Hugepages — memory breakdown (total, available, buffers, cached) plus 2MB/1GB hugepage stats. Allocate Hugepages picks page size and count and applies immediately.
  4. Optimization — per-VM recommendations (resource, current → recommended, reason, impact). Apply runs the change against that VM.

Typical flow: learn topology → allocate hugepages if needed → open Optimization → Apply carefully on non-prod first → confirm on VM detail / Live Metrics. For live utilization (not layout), use System Health.

Empty / fail: Error banner, empty table, or failed toast — confirm you are signed in, zyvor-fabricd is healthy (/readyz on http://127.0.0.1:<port> or https://<host>), and any backend this page needs (FluxVM, storage, network) is reachable. See Admin basics.

Success: Live data loads without error; creates/updates appear in the list or detail view and any confirmation toast clears cleanly.