QueueWorker for failed pipelines

Problem/Motivation

As there currently is a server problem with the resmush.it servers, images don't get optimized during derivative creation, but the derivatives get saved in an unoptimized/failed state. Therefore it would be nice, if the imageapi_optimize module would track these failed pipeline processor calls and retry to execute them later again. processor plugins already inform the optimization whether they succeeded, so this return value can be used to add a failing image to a queue to reprocess it later again (e.g. during cron runs)

Proposed resolution

• Create a queue to track failed image pipeline processes
• Create QueueWorker to run during cron runs to reprocess failed image pipelines
• Create a drush task to process the queue
• ... and perhaps this retry behavior should be configurable (enabled/disabled) per pipeline (e.g. to avoid retries of corrupt images when using local binaries for optimization, where you may be sure that it is not a problem with an external service

Remaining tasks

• Create a patch that introduces this new functionality

User interface changes

n/a

API changes

n/a - Processor plugins only have to ensure that they return the correct state within implementations of ImageAPIOptimizeProcessorInterface::applyToImage()

Data model changes