Skip to main content

Handling Failures

The processors in a pipeline run sequentially, and when one of them fails or encounters an error, by default the process halts.

The on_failure processor can be used to specify a list of processors to run immediately after a processor fails. However, Director keeps running the pipeline's remaining processors even if the on_failure field is blank.

Recovery​

To ignore a failure and keep running the remaining processors in the pipeline, ignore_failure must be set to true.

- rename:
    field: provider
    target_field: cloud.provider
    ignore_failure: true

Continuation​

To configure a processor so that in the event of a failure another processor runs, the on_failure parameter can be used.

- rename:
  field: foo
  target_field: bar
  on_failure: [
    - set:
        field: error.message
        value: "Field 'foo' could not be renamed as 'bar'"
        override: false

Nesting​

For cascading errors, multiple on_failure processors can be nested.

- rename:
    field: foo
    target_field: bar
    on_failure:
      - set:
          field: error.message
          value: "Field 'foo' could not be renamed as 'bar'"
          override: false
          on_failure:
            - set:
              field: error.message.cascade
              value: "Multiple errors encountered"
              override: true

Pipelines​

The on_failure can also be specified for the pipeline. When a processor fails, in the absence of a processor-level on_failure setting, Director uses the pipeline-level setting as a fallback. However, in that case it does not attempt to run the remaining processors.

on_failure:
  - set:
      field: _index
      value: failed-{{{ _ingest.on_failure_processor_tag }}}
      message: "Processor failed"

Metadata​

More information may be available in the metadata fields on_failure_message, on_failure_processor_type, on_failure_processor_tag, and on_failure_pipeline which can only be accessed from within an on_failure block.

on_failure:
  - set:
      field: error_info
      value: "{{ _ingest.on_failure_processor_type }} in {{ _ingest.on_failure_pipeline }} failed; error message: {{ _ingest.on_failure_message }}"