Skip to content

Kuiper 0.9.0 version

Pre-release
Pre-release
Compare
Choose a tag to compare
@jinfahua jinfahua released this 06 Aug 01:09
· 2381 commits to master since this release
0.9.0
59f718a
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

Overview

Kuiper 0.9.0 is a major version that includes the stream state management, KubeEdge device model adoption, EdgeX array type support and TDengine database sink support.

Features & fixes

  • State management function is supported. This feature enables Kuiper to implement stream with state:
    • Fault tolerant processing is supported. If an unexpected interruption occurs during stream processing, it can be resumed after the rule is restarted;
    • Checkpointing is supported, which allows users to implement QoS settings during stream processing, including At-most-once(0), At-least-once(1) and Exactly-once(2);
    • By consuming data from the specified offset, users can extend the relevant interface to realize the data source that can re-consume the data, so that it can be recovered offline or when there is an error in stream processing;
    • Configurable state persistent storage. The system stores the state in the file system by default, and also supports storing state data in a third-party system, such as Redis;
    • The user is supported to call the API provided by Kuiper to realize the storage of custom state data when extending the source, sink and function;
  • Access support for KubeEdge data model is provided, with a tool for automatically issuing configuration files through Kubernetes configmap. Users can use Kuiper to directly support the analysis of device data from KubeEdge;
  • TDengine plug-in is added, which can support saving analysis results to TDengine time series database;
  • All Chinese documents are translated and synchronized;
  • The execution process of FVT in Github Action is optimized, and some unnecessary test processes are deleted;
  • filter in the window with support for filtering data can realize the data filtering and window grouping; this function is more important for counting windows: the result of filtering through the WHERE statement and then window grouping will be different;

Thanks

Assets 42
Loading