Elastic exceptions may not lead to a failure exit status

[jdnurmi]

Use case:

• Emails are delivered to a shared filesystem available to a pod running (similar to a Maildir)
• For each file, call parsedmarc - which is configured to emit reports to ElasticSearch, and on exit 0, delete the file.

Problem:
Parsedmarc doesn't always exit non-zero on failure, even when its only configuration is to write reports.

Traceback led me to cli.py:133, and it looks like exception swallowing happens a lot here - so I didn't know what sort of patch might be accepted upstream.

For now I wrapped the whole thing in a grep for 'Elasticsearch (exception|Error)' to determine if it was successful - but figured I'd open this up for discussion -- but also my use-case is weird, so it may not be applicable to many other people.

[nhairs]

Hi @jdnurmi,

You may be interested in an experimental fork I've been working on which does have a feedback mechanism for failed processing by Elasticsearch (or any of the destinations) so that the source of the reports can appropriately handle them.

nhairs/parsedmarc-fork