[FLINK-31508] FLIP-304: Pluggable failure handling for Apache Flink

• Type: New Feature
• Status: In Progress
• Priority: Major
• Resolution: Unresolved
• Fix Version/s: 2.0.0
• Component/s: Runtime / Coordination, Runtime / REST
• Assignee: Panagiotis Garefalakis
• Reporter: Panagiotis Garefalakis
• Created: 17/Mar/23 19:25
• Updated: 10/Jul/24 07:03

Description

This is an umbrella ticket for FLIP-304

Issue Links

• links to: FLIP-304

Sub-Tasks

• Introduce interfaces and utility classes related to enrichment/labelling of failures leading to job restart: Resolved
• Add documentation for implementing/loading enrichers: Closed
• Translate failure-enricher documentation to Chinese: Resolved
• Introduce DefaultScheduler failure enrichment/labeling: Resolved
• Introduce AdaptiveScheduler per-task failure enrichment/labeling: Resolved
• Introduce AdaptiveScheduler global failure enrichment/labeling: Resolved
• Introduce AdaptiveBatchScheduler failure enrichment/labeling: Resolved
• Initialize and pass down FailureEnrichers: Resolved
• ExceptionHistory and REST API failure label integration: Resolved
• End-to-end integration tests for failure labels: Resolved
• Make FailureEnricherContext more flexible and include task specific fields: Open
• Extend web interface to support failure labels: Resolved
• Log an error when enrichers defined as part of the configuration can not be found/loaded: Resolved
• Unify the GlobalFailureHandler and LabeledGlobalFailureHandler interface: In Progress
• Protect failure enrichment against unhandled exceptions: Closed