The overall "Enclosure Status" from getencstat latches on to "CRITICAL" if the status changes. (E.g., if one power supply loses power for a couple seconds.) Reboots do not clear this status, so it stays latched until it is manually cleared with setencstat.
We don't alert on it, but it does show up in debugs forever. (Approximately half of our stored debugs show "/dev/ses0: Enclosure Status <CRITICAL>")
The status can be set back to OK with "setencstat /dev/sesX 0".
When we see this status, I think we should log it and clear it, but not alert on it. I think that would give us an approximate time of when the event happened, and would leave better output in debugs.
This is not critically important as I'm pretty sure that this particular field does not get interpreted by our enclosure page. But customers do see the getencstat output sometimes, though not often.