Architecture
How the QuilrAI Process Mapping engine discovers, monitors, and governs every application and AI agent on the endpoint.
Endpoint
# Discovery sources
Process Monitor → polls every 10s (sysinfo)
File Scanner → startup + 30min poll (Lua sandbox)
Backend SSE → governance overrides in real-timeQuilrAI Process Mapping
Discover
Process Monitor
sysinfo polling (10s)
PID reuse detection
New / exited process tracking
File Scanner
Lua scripting sandbox
AI config file discovery
Startup + 30 min poll
Correlate
Correlator
Two-level cache (alias + app)
300s TTL per entry
Path-based resolution
EntityStore
DashMap (concurrent)
Broadcast change events
Added / Updated / Removed
Sync
Ingest Service
POST /sync/discovered-apps
Gzip-compressed batches
Canonical ID assignment
Delta Sync
GET /sync/delta
Governance overrides
Approval status + policy
Enforce
Enforcer
Block: kill + notify
Quarantine: kill + log
Allow / Audit: log only
Snapshot Timer
JSON dump every 30s
Crash recovery on startup
Idempotent delta replay
Quilr Backend
App CatalogGovernance PoliciesProcess MappingsDelta EndpointAudit DashboardSSE Stream
QuilrAI
Pipeline Stages
Every discovery and process event flows through these stages in order.
Entity Types
Persistence
In-memory DashMap for concurrent hot-path access, with periodic JSON snapshots for crash recovery.
Platform Support
Observability
Every discovery event, governance override, and enforcement action is logged with entity identity, policy match, and action taken. Ingest Service and Delta Sync maintain continuous bidirectional sync with the Quilr backend.