Skip to content
This repository has been archived by the owner on Aug 2, 2022. It is now read-only.

Initial Release - Version 1.7.0.0
Compare
Choose a tag to compare
@kaituo kaituo released this 04 May 23:41
v1.7.0.0
865f5fc
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.

Compatible with Elasticsearch 7.6.1 and Open Distro for Elasticsearch 1.7.0.

The Open Distro for Elasticsearch Anomaly Detection plugin enables you to detect anomalies in streaming time series data based on the random cut forest (RCF) algorithm.

You can create anomaly detectors and add features to them to customize what data you want to detect anomalies over.

This plugin can be used with the Anomaly Detection Kibana plugin for an intuitive user interface that can be used to configure, start, and stop anomaly detectors. You can also view the anomaly history for all of your created detectors.

This plugin works independently. You can also use the plugin with the same version of the Open Distro for Elasticsearch Alerting Kibana plugin to get alert notifications. You can create a monitor based on an anomaly detector directly on the Alerting Kibana plugin. Monitors run checks on the anomaly detection results regularly and trigger alerts based on custom trigger conditions.

Features

  1. Create and configure anomaly detectors over user-specified indices and features
  2. Start and stop detectors at any time
  3. Query anomaly results
  4. Query existing detectors
  5. Query specific detector details and current state
  6. Preview anomaly results for new detector features

Current Limitations

  • Limit of 1000 detectors per Elasticsearch cluster
  • Limit of 5 features per detector
  • Total detectors memory limit of 10% of JVM heap
  • Not all API calls have complete error handling
  • We will continuously add new unit test cases, but we don't have 100% unit test coverage for now. This is a great area for developers from the community to contribute and help improve test coverage
  • Please see documentation links and GitHub issues for other details

Major Changes

  • Add state and error to profile API PR #84
  • Preview detector on the fly PR #72
  • Cancel query if given detector already have one PR #54
  • Support return AD job when get detector PR #50
  • Add AD job on top of JobScheduler PR #44
  • Adding negative cache to throttle extra request PR #40
  • Add window delay support PR #24
  • Circuit breaker PR #10 PR #7
  • Stats collection PR #8

Enhancements

  • Stats API: moved detector count call outside transport layer and make asynchronous PR #108
  • Change AD result index rollover setting PR #100
  • Add async maintenance PR #94
  • Add async stopModel PR #93
  • Add timestamp to async putModelCheckpoint PR #92
  • Add async clear PR #91
  • Use callbacks and bug fix PR #83
  • Add async trainModel PR #81
  • Add async getColdStartData PR #80
  • Change the default value of lastUpdateTime PR #77
  • Add async getThresholdingResult PR #70
  • Add async getRcfResult PR #69
  • Fix rcf random seed in preview PR #68
  • Fix empty preview result due to insufficient sample PR #65
  • Add async CheckpointDao methods. PR #62
  • Record error and execution start/end time in AD result; handle except… PR #59
  • Improve log message when we cannot get anomaly result PR #58
  • Write detection code path in callbacks PR #48
  • Send back error response when failing to stop detector PR #45
  • Adjust preview configuration for more data PR #39
  • Refactor using ClientUtil PR #32
  • Return empty preview results on failure PR #31
  • Allow non-negative window delay PR #30
  • Return no data error message to preview PR #29
  • Change AD setting name PR #26
  • Add async CheckpointDao methods. PR #17
  • Add async implementation of getFeaturesForSampledPeriods. PR #16
  • Add async implementation of getFeaturesForPeriod. PR #15
  • Add test evaluating anomaly results PR #13

Bug Fixes

  • Change setting name so that rpm/deb has the same name as zip PR #109
  • Can't start AD job if detector has no feature PR #76
  • Fix null pointer exception during preview PR #74
  • Add threadpool prefix and change threadpool name PR #56
  • Change setting name and fix stop AD request PR #41
  • Revert "merge changes from alpha branch: change setting name and fix … PR #38
  • Fix stop detector api to use correct request PR #25

Infra Changes

  • Add release notes for ODFE 1.7.0 PR #120 PR #119
  • Open Distro Release 1.7.0 PR #106
  • Create opendistro-elasticsearch-anomaly-detection.release-notes.md PR #103
  • Update test branch PR #101
  • Bump opendistroVersion to 1.6.1 PR #99
  • Change to mention we support only JDK 13 PR #98
  • AD opendistro 1.6 support PR #87
  • Added URL for jb_scheduler-plugin_zip instead of local file path PR #82
  • Change build instruction for JDK PR #61
  • ODFE 1.4.0 PR #43
  • Add spotless for code format PR #22
  • Update third-party PR #14
  • Build artifacts for rpm, deb, zip PR #5
  • Update test-workflow.yml PR #2
Assets 2
Loading