Windower Streaming Transformation

Plugin version: 2.11.0

The Window plugin is used to window a part of a streaming pipeline.

Use Case

Windows are used when you wish to group smaller batches of events into larger batches of events. For example, if your streaming pipeline has a batch interval of one second, you may want to accumulate thirty seconds of events before running an aggregation, in order to accumulate enough events to generate useful aggregates. You may also want to slide the window at a slower frequency. For example, you may want to calculate those aggregates over every five seconds instead of every second.

Configuration

Property	Macro Enabled?	Description

Property	Macro Enabled?	Description
Width	No	Required. The width of the window in seconds. Must be a multiple of the batch interval. Each window generated will contain all the events from this many seconds.
Slide Interval	No	Required. The sliding interval of the window in seconds. Must be a multiple of the batch interval. Every this number of seconds, a new window will be generated.

Example

This example creates windows of width five seconds that slides every two seconds. In order to use this configuration, the batch interval for the entire pipeline must be set to one second, as both the width and slide interval must be multiples of the batch interval.

Property	Value

Property	Value
Width	`5`
Slide Interval	`2`

With this setup, the plugin would generate these windowed batches:

time	input	output

time	input	output
1	x1
2	x2	[ - - - x1 x2 ]
3	x3
4	x4	[ - x1 x2 x3 x4 ]
5	x5
6	x6	[ x2 x3 x4 x5 x6 ]
7	x7
8	x8	[ x4 x5 x6 x7 x8 ]
9	x9
10	x10	[ x6 x7 x8 x9 x10 ]

2.4.3