We use a Saleae logic probe to catch seldom seen erratic behaviour. This involves often days of unattended operation in "wait for trigger" mode. This morning, I found a "couldn't get pace on interface" error message. I don't know why. The sample rate isn't that close to the edge. I suppose that there was some transient action of another task in Windows. After a restart (same sample rate), it runs for hours now, like it did the weeks before.
Hence my two proposals: When the error message pops up, show date and time of day when the error occured.
Enable an optional recover mode. When pace is lost, recover automatically and continue waiting for trigger event, whereas the fault event is logged to a log. So, in the morning one can see whether there was a "blind period" during the night and, in case, when.