I have a project that heavily relies on queue processing. It is deployed as a Windows Docker container and hosted on a Windows Server on the client’s premises.
Today, the client reported that some queues were not being processed correctly. Some queue entries randomly failed with errors like “Unable to load xxx” or “The type initializer for ‘yyy’ threw an exception.” When I forced another attempt to process the items, some of them were processed successfully and most of them failed. However, the number of failed items was steadily decreasing.
Also, when I processed the entries manually, the process worked correctly.
Strangely enough, these errors were missing from the logs despite the log level being set to ERROR.
There were no recent server updates or changes to the application.
In the end, a complete restart of the host machine resolved the situation, but any insight into what could have caused such behavior would be appreciated.