Overview vPW Architecture

Components

• analyzer
  • Spring-Boot-App
  • Provision of a REST API for the vPW frontend
  • Persistence of inventory data (dashboards, indicators) in PostgreSQL
  • Generation of Elasticsearch requests from the frontend requests and processing/preparation of the results
• pipeline
  • Spring-Boot-App
  • Receiving process data from Kafka
  • Data preparation, including “breaking up” JSON objects into flat structures
  • Storage/indexing in Elasticsearch
• vpw-shared-elasticsearch-config
  • Spring boot AutoConfiguration for the Elasticsearch REST client, used by Analyzer and Pipeline.
• web-app
  • Frontend, user interface & dashboard
  • HTML/CSS/Javascript
  • NGINX webserver and reverse proxy
  • Communication with pipeline and analyzer via REST (JSON)
• polling-client
  • extraction of Camunda process data and their permanent storage in Apache Kafka
• elasticsearch
  • Central vPW data store and basis for evaluations
  • Simple data model: Process, Activity, Variable (Parent/Child relationships)
  • Key figure calculation “on-the-fly” with ad-hoc queries (filters, aggregations)
• kibana
  • Visualization of Elasticsearch data as an addition to the vPW frontend (web-app)

Characteristics

• Asynchronous processing
  • No influence on normal operation
• Open architecture
  • possible integration of different event sources
  • connection of own clients to Kafka and Elasticsearch
• “Near-real-time” processing
• Scalability and stateless components
• Operation on container platforms