I am seeing a number of reports in many different threads since the March meltdown and the subsequent implementation of the new scheduler indicating random problems with a number of different zigbee devices, including smart bulbs, motion sensors, pocket sockets, etc.
Most people are reporting this as a failure of the device, but I’m not sure it is. It may be a failure of the cloud to properly record the events, or database corruption, or local flooding.
It may be devices using a device handler that has the known millisecond problem. The new scheduler which was rolled out in March has a missing converter and treats milliseconds as seconds, so device handlers with programmed delays will wait 1,000 times longer than they should. This might cause a timeout, or what appears to be a failure to respond that then results in a mysterious execution some time later. Again, not a device failure.
For example, people are reporting that zigbee bulbs are not able to be controlled from SmartThings, but if they rename the device, then it works again. OK, that’s obviously not a device failure.
Other people, including me, are seeing zigbee bulbs (including Cree and GE) randomly turn on. Dropping off might be a device failure (the GEs are known for this) but randomly turning on isn’t, usually.
Several people, including me, Have zigbee motion sensors which appear to be being ignored by SmartThings. Or zigbee contact sensors reporting only open or only closed for a period of time – – and then fixing themselves. Granted, there are many reasons why sensors might not be reporting correctly (we have an FAQ for that) but these are devices that were working just fine, then seem to stop working for some random period of time, then start working again. Yes some of this gets disguised because many people would just go ahead and pop The batteries on the sensors immediately. I don’t do that because I have to wait for someone else to change the batteries (limited hand function), so I see them come back.
Reason I’m posting this now is there was a really bizarre incident here yesterday. At about six in the morning two of our five zigbee sensors, both smart things branded motion sensors, began to be ignored by smartthings for motion. However, their temperature reports continued to be processed all day. This continued until this morning, when both reports were again processed. As it happens, we didn’t do anything with the sensors in between except to verify that the reports are being ignored. We didn’t move them, change the batteries, rename them, re-– pair them, nothing.
If this were a controller failure, I would expect all the zigbee devices in the Home to be affected. I’m quite sure it’s not a device failure, particularly since two showed exactly the same fault behavior at the same time and they are in different parts of the house. And others zigbee devices in the same area continued to work fine. And both sensors came back at about the same time. 
So it seems like it must be some kind of Cloud issue.
Anyway, if you are seeing weird things with Zigbee devices, I would encourage you to Report it to support. I suspect this kind of problem would often be misidentified as a device issue, and therefore support would not get as many notifications as they would for schedule failures. So even if there is a cloud problem going on, they might not be aware of it.
But that’s just me.