Was the stopping_timeout value changed for the affected appliances? I believe some appliances were modified from the default of 10 minutes. Can you provide the logs including the log/last_settings.txt? If not: 1) What is the value for stopping_timeout? 2) What is the memory_threshold for the "MiqReportingWorker" and "MiqSmartProxyWorker"? It appears we're very frequently breaking the memory_threshold. We should break this threshold infrequently. We need to raise this threshold based upon how much memory is required. We currently have a default of 500.megabytes for the reporting worker and 600.megabytes for the smart proxy worker, so that is the minimum. In order to not have work go uncompleted, we should raise the threshold to values which enables these workers to complete a majority of work without tripping it. We can then determine if there are ways to make that work use less memory.