apache / apache/beam

Give a option to ignore the timer that is larger than END_OF_GLOBAL_WINDOW

Open
#18,443 0 comments 0 reactions 0 assignees View on GitHub
flink improvement P3 runners
Dominant language
Java
Stars
8.7k
Forks
4.7k
Avg merge
2d 5h
Merged PRs (30d)
204

Description

Event time can not reach END_OF_GLOBAL_WINDOW in unbounded world. (Except for testing)
But Flink runner will set some timers when user set a StatefulPardo/GBK with GlobalWindow. Flink maintains timers in PriorityQueue on the Java Heap. There is a bad performance when the number of keys is very much.

Imported from Jira [BEAM-2487](https://issues.apache.org/jira/browse/BEAM-2487). Original Jira may contain additional context.
Reported by: lzljs3620320.

Contributor guide

Open the contributing guide

Research direction

Start by tracing the Flink runner's timer handling for StatefulPardo/GBK with GlobalWindow and how timers beyond END_OF_GLOBAL_WINDOW enter the PriorityQueue. Done means the runner provides the requested way to ignore those timers and avoids retaining them for the affected keys.

Written by the indexing model from the issue text.

Assessment

Tech stack
java
Domain
stream-processing
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.