Skip to main content

Progressive Deployment

Watchlight Beacon uses a three-tier deployment model. Start with visibility into your AI infrastructure, then layer on authorization and runtime governance as your needs grow.

Tier 1: Discovery & Registry

"What agents and MCP servers are running in my environment?"

ServicePurpose
wl-registryAgent & MCP server catalog
wl-registry-frontendRegistry dashboard
wl-discoverNetwork scanner + agent detector
PostgreSQLDatabase
OpenBaoSecret management
CaddyTLS reverse proxy

What you get:

  • Automatic discovery of MCP servers and AI agents across your infrastructure
  • Centralized registry with trust state management (unverified / trusted / quarantined / revoked)
  • Docker container security metadata collection and risk signals
  • Multi-tier agent framework detection (LangGraph, CrewAI, AutoGen, LangChain)
  • Interactive topology visualization

Requirements: Docker Engine 24+, 2 GB RAM, port 8443

Tier 2: Authorization & Policy

"Who can do what? Enforce least-privilege policies on agent actions."

Everything from Tier 1, plus:

ServicePurpose
wl-apdpCedar policy authorization engine
wl-apdp-frontendPolicy management dashboard

What you get (in addition to Tier 1):

  • Intent-based authorization with Cedar Policy Language
  • Goal management — time-boxed, action-limited business objectives
  • Delegation chains — cryptographically-verifiable trust paths
  • Intelligent policy selection (20-30x performance improvement)
  • Observation mode (default) or enforcement mode
  • Shared PostgreSQL database — agents discovered by wl-discover are immediately available for policy evaluation

Requirements: Docker Engine 24+, 4 GB RAM, ports 443 + 8443

Tier 3: Runtime Governance

"Transparently enforce policy on every API call. Agents never see raw API keys."

Everything from Tier 2, plus:

ServicePurpose
wl-proxyGovernance proxy — policy eval + credential injection
wl-secrets-brokerPolicy-enforced secrets management (SCT issuance)

What you get (in addition to Tier 2):

  • Transparent governance — every API call evaluated by Cedar policies before execution
  • Credential injection — agents never see raw API keys; wl-secrets-broker injects credentials
  • Secret Capability Tokens (SCTs) — Ed25519-signed, time-limited, single-use tokens
  • Full audit trail — every secret access, API call, and policy decision logged
  • Zero-configuration agent updates — change 1-2 environment variables, no code changes

Requirements: Docker Engine 24+, 8 GB RAM, ports 443 + 8443 + 9443

Choosing Your Starting Point

Most organizations start with Tier 1 to gain visibility, then upgrade as their governance needs mature. Each tier builds on the previous — upgrading means running a new bootstrap script that includes all prior services.

Tier 1 ──────▶ Tier 2 ──────▶ Tier 3
Visibility Authorization Runtime Governance

Next Steps