When you configure "Auto-Update" for a tag, the server will send a command to the tag to configure it to send periodic status update (with temperature etc) at specified interval. An "out of range" notification is sent when the server does not receive periodic status update from a tag at expected time + some margin.

Because the tag stores the interval configuration in its local volatile memory, if you remove the battery and reinsert it, the tag will forget about the periodic status update it is supposed to send. As a result, soon you will receive a false "out-of-range" notification even though the tag is well within range. When you "ping" the tag, it will come back in range right after. 

When the tag tries to transmit status update the transmission fails for some reason, it thinks it is out of range and will enter into a "hibernate" mode, where it no longer sends any status update until receiving a new command from the Tag Manager. If the status update fails for enough number of times to make the tag enter hibernate mode, and then recovers afterwards, the tag does not know about it and soon you will receive a false "out-of-range" message. 

We have made a simple yet effective change to the server-side algorithm to reduce these false out-of-range notifications: Just before the server is about to mark the tag "out-of-range", it automatically attempts to re-configure the tag's "Auto-Update" interval, and only marks the tag out of range when this fails. Now when you receive "out-of-range" notification you know with more certainty that the tag is really out of range. Because this change is done on the server side, no app date / firmware update is needed.