Skip to main content

Dashboard Overview

Feature rows 52–53 — Sprints UX1 / UX1.1 / UX1.2

The dashboard is included on every plan.

The shape

The Behavry dashboard is structured around three persistent regions:

  • Sidebar — a 52-pixel icon-only rail with 10 hub icons. Each icon is a page in its own right (not just a category). Clicking an icon navigates to the hub; inside the hub, tabs switch between subviews.
  • Topbar — a greeting message, a role chip, a PeriodSwitch for "data from the last 7 / 30 / 90 days," and an EscalationButton that lights up when there's a pending HITL item.
  • StatusSpine — a thin horizontal line beneath the topbar that acts as the tenant-wide heartbeat. Green when everything is nominal, amber or red when one of the core subsystems needs attention. Hovering shows a breakdown.

The 10 hubs

IconHubContents
OverviewOverviewHome + Dashboards (configurable widget grid)
AgentsAgentsList, Sessions, Workflows, Provisioning
ActivityActivityEvents, Escalations
BehaviorBehaviorMonitoring, Alerts
PoliciesPoliciesLibrary, DLP, Inbound Rules, Context Gate, Suggestions, Changes
LibraryLibraryCommunity browse / subscribe / install
DiscoveryDiscoveryAI Surface Map, Dependency Graph, Citizen Coder
ObservabilityObservabilityMetrics, Cost, Logs
ComplianceComplianceFramework mappings, evidence, export
SettingsSettingsThe Settings Hub

The 10-hub structure replaced the older monolithic sidebar; 16 backwards-compatible redirects keep old URLs working.

PeriodSwitch

Every time-scoped view on the dashboard (cost, behavioral baselines, compliance status, etc.) respects the PeriodSwitch in the topbar. Options: 24h / 7d / 30d / 90d. Your choice is persisted per-user.

StatusSpine

The StatusSpine is the fastest way to answer "is everything OK right now?" It aggregates six core signals:

  1. Identity — any failed JWT validations in the last 5 minutes
  2. Policy — any OPA compile failures or policy errors
  3. Proxy — any 5xx from the proxy pipeline
  4. Audit — any audit write lag above threshold
  5. SIEM — any destination with a growing dead-letter queue
  6. HITL — any escalation waiting longer than its SLA

Green means everything's nominal. Amber means one signal is degraded. Red means at least one is failing. Hovering shows the per-signal breakdown and deep-links to the relevant hub.

Greeting & role chip

The greeting in the topbar uses the authenticated user's first name (or username fallback) and shifts with time of day. Next to it, a role chip shows the active role: viewer, analyst, policy_author, admin, or super_admin. Super-admins also see the active tenant in a second chip.

Per-user layout

Most hubs are static — a consistent list that isn't user-configurable. The Overview → Dashboards tab is the exception: it's a Widget Dashboard where every user gets their own grid. Super-admins can publish persona templates (CIO / CTO / CISO) that new users can apply as a starting point.