Documentation
Docs Home › Observability › Manager Visibility

Manager Visibility

Purpose

List the Manager console and API surfaces operators use to see fleet health, eBPF attach, traffic widgets, and Hosting evidence — without treating the WebSocket as a public stream product.

Audience

Administrators operating the SPA at /login and zwallctl.

Edition

SHARED dashboard groups. GET /api/v1/dashboard/ebpf-health is Enterprise in the v1 router. Hosting pages are Enterprise. Some traffic history endpoints are license-gated as advanced reporting.

Status

SUPPORTED as implemented widgets. Field-level schemas are not a second public OpenAPI contract.

Prerequisites

Ready Manager (GET /health/ready), admin session, Agents heartbeating.

Required permissions

Admin for the routes below. Viewer is insufficient for most operator APIs.

Dependencies

PostgreSQL; Agent heartbeats (agents.timeout example 2m marks offline); optional Hosting or Virtualization Agents for those widgets; web.enabled: true for the SPA.

Configuration

CORS server.cors.allowed_origins must include the browser origin. Metrics widgets on the dashboard are distinct from the Prometheus scrape at /metrics.

Procedure

Console after login at /login. Corresponding admin APIs:

Scroll sideways to see the full table. Use the arrow keys when focused.

SurfacePath
Dashboard statsGET /api/v1/dashboard/stats
Dashboard agentsGET /api/v1/dashboard/agents
eBPF healthGET /api/v1/dashboard/ebpf-health (Enterprise)
Logs widgetGET /api/v1/dashboard/logs
System widgetGET /api/v1/dashboard/system
Dashboard metricsGET /api/v1/dashboard/metrics
TrafficGET /api/v1/dashboard/traffic and /api/v1/traffic/* (some history license-gated)
Security events widgetGET /api/v1/dashboard/security-events
Live console updatesGET /api/v1/dashboard/ws
SanityGET /api/v1/system/sanity-check
HostingGET /api/v1/agents/:id/hosting/status and /hosting/events
VirtualizationGET /api/v1/virtualization/vms and /interfaces
FleetGET /api/v1/agents

CLI: zwallctl list-agents, list-events, log, stats, check, ping, status --agent ID.

zwallctl log --follow prints not implemented. zwallctl stats reads local eBPF maps (stats_map) on a Virtualization host with pin dir /sys/fs/bpf/fw — it is not a Manager dashboard substitute.

Verification

Dashboard loads; zwallctl check / sanity-check returns without auth errors; Hosting and virtualization pages match installed profiles.

Expected result

An operations console view of Agents, attach health, and recent events. Not a SIEM and not a public streaming API.

Failure modes

Scroll sideways to see the full table. Use the arrow keys when focused.

SymptomAction
Empty eBPF healthEnterprise-only route or no Virtualization Agent
Empty Hosting UIno Hosting heartbeat; see source states
WS failssession cookie / HTTPS; cookie issuance requires acceptable HTTPS
Traffic history 403license gate — not a broken Agent
CLI default :8080set --manager-url

Troubleshooting

Diagnostics, Console. Confirm /health/ready before debugging widgets.

Security considerations

WebSocket is admin-equivalent. Do not expose the Manager UI without TLS. Treat /metrics scrape tokens as admin (see Prometheus and Grafana).

Limitations

Not Splunk. Cluster Grafana dashboard is not a public HA product. OpenAPI is not a generated customer contract.