bamaer opened a new pull request, #9427: URL: https://github.com/apache/hop/pull/9427
When a pipeline, workflow, transform or action fails, Hop now builds an `ExecutionFailure`. It says what failed, in which phase, why (exception, root cause, SQL state) and in which category, with credentials redacted. - Extension points `TransformFailed`, `PipelineFailed`, `ActionFailed`, `WorkflowFailed`, plus `addExecutionFailureListener()` on the local engines. - `ExecutionState.failure` at every execution information location. A pipeline that fails to prepare or initialize is now registered there too. - Hop Server pipeline and workflow status: `errorDescription` is filled, and the json form has the `failure` object. - `hop-run --result-json <file>` (opt-in). Points worth checking: - **Mixed versions:** Hop 2.20 and earlier can't read an Elastic or OpenSearch document of a failed run written by this version, and then show that location as not reachable. Documents of runs without a failure are unchanged. This version reads leniently. - **Existing Elastic and OpenSearch indexes:** the location adds an unindexed `executionState.failure` mapping before its first write. If that's refused (an API key that can only write documents, for example), the failure is mapped field by field. - **Transform state:** when `init()` throws, the transform's execution state now shows `failed`, like the pipeline already did. Error counts are unchanged. - Only the local engines report failures. Remote, Beam and Spark runs have none. Tests: unit tests in core, engine, elastic, opensearch and neo4j. ITs: `hop_run` main-0005 (result file, and the failure chain through an Abort on a failure hop) and `opensearch` main-0002 (failure stored and mapped unindexed). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
