@startuml insight_quota_flow
!option handwritten true
skinparam sequenceMessageAlign center
skinparam defaultFontName Helvetica
autonumber

participant "DB" as DB
participant "event-server" as EventServer
participant "insight-task-generator" as TaskGenerator
participant "insight-task-processor" as TaskProcessor
participant "vlm-host" as VlmHost

EventServer -> TaskGenerator : Webhook Event (Trigger)
TaskGenerator -> TaskProcessor : Send Task Message
TaskProcessor -> VlmHost : Call VLM Model
VlmHost --> TaskProcessor : Model Results
TaskProcessor -> TaskProcessor : Process Results
TaskProcessor -> DB : Save InsightEvent

@enduml