A pipeline can encounter failure for two primary reasons: either due to an individual node within the pipeline encountering an error during execution, or because a edge definition in the pipeline failed. Here’s how to troubleshoot and understand pipeline failures:
Node-Level Failures:
- Each node in a pipeline represents an execution, task, or deployment.
- To investigate node-level failures, you can click on the pipeline graph to select any specific node of interest.
- Additionally, you can review the logs associated with each node to identify errors or issues during execution.
Pipeline-Level Errors:
- Pipeline-level errors, such as issues arising from the pipeline’s overall structure or configuration, can be found under the “Log” tab within the pipeline view.
Common Error Messages
-
“Node ‘name’ transitioned to crashed”: This message indicates that one of the executions within the pipeline has failed. To resolve this:
- Check the pipeline’s executions to identify the one that has completed with an error.
- Examine the logs associated with this specific execution to understand the cause of the failure and take corrective action.
-
“Stopping due to 1 incompletable edges”: This error suggests that a node couldn’t initiate because it lacked necessary inputs required for that particular step.
- Review your pipeline edges carefully to ensure that all the required files and parameters are correctly passed between nodes.
- Verify that the data flow within your pipeline is well-defined and all dependencies are met.
By investigating node and pipeline-level errors and addressing them accordingly, you can diagnose and resolve issues in your pipelines to ensure successful execution.
