Skip to main content

Window Investigation Screen

Value and Problem Solved​

The "Window Investigation Screen" enables users to deeply analyze and investigate the behavior of specific windows in their Flink jobs. This feature helps users:

  • Investigate a window in depth to gain a comprehensive understanding of its performance and behavior.
  • View detailed data on records, including the distribution of their arrival times.
  • Seamlessly navigate to other windows or related features for further inquiries and debugging.

By providing these capabilities, the feature empowers users to identify issues, optimize pipelines, and ensure efficient data processing.

Functionality and How It Works​

Graph View​

The graph view allows users to visualize the performance and behavior of a specific window. Users can select what data to display in the graph, including:

  • Records on time: Displays the number of records that arrived on time.
  • Late arrivals: Shows the number of late-arriving records.
  • Ideal throughput line: Indicates the expected throughput for optimal performance.
  • Watermark threshold: Displays the watermark values for the selected window.
  • Window emits: Shows the emit events of the window.

This interactive graph enables users to correlate performance metrics with their visual representation for better understanding and optimization.

Tables View​

The tables provide a detailed breakdown of input records and emits within the selected window:

  • Input Records Table: Lists all records in the window, categorized into on-time and late records.
  • Emits Table: Displays all emits of the window with timestamps and other relevant details.

This tabular representation offers users a structured and clear overview of the data for deeper analysis.

Navigation to Other Screens​

Users can navigate to other relevant screens with the appropriate context from the window investigation screen. This enables them to:

  • Investigate related metrics or issues in other parts of the Flink pipeline.
  • Maintain continuity in debugging by preserving the context of the current investigation.

Interaction with the Timeline​

  • Users can drag and drop the timeline touchpoint using a brush to navigate to checkpoints
  • All graphs and logs will updates accordingly.