Looks to be Interference and Noise exceeded:
ARM History
-----------
Time of Change Old Channel New Channel Old Power New Power Reason
-------------- ----------- ----------- --------- --------- ------
2013-07-09 14:51:39 1 11 21 21 N
2013-07-09 14:51:38 11 1 21 21 I
2013-07-09 14:32:39 1 11 21 21 N
2013-07-09 14:32:38 11 1 21 21 I
2013-07-09 14:21:46 1 11 21 21 N
2013-07-09 14:21:45 11 1 21 21 I
2013-07-09 14:15:29 1 11 21 21 N
2013-07-09 14:15:28 11 1 21 21 I
2013-07-09 14:11:05 1 11 21 21 N
2013-07-09 14:11:04 11 1 21 21 I
2013-07-09 14:06:54 1 11 21 21 N
2013-07-09 14:06:53 11 1 21 21 I
2013-07-09 13:59:04 1 11 21 21 N
2013-07-09 13:59:03 11 1 21 21 I
2013-07-09 13:48:13 1 11 21 21 N
2013-07-09 13:48:12 11 1 21 21 I
2013-07-09 13:11:12 1 11 21 21 N
2013-07-09 13:11:11 11 1 21 21 I
2013-07-09 13:06:52 1 11 21 21 N
2013-07-09 13:06:51 11 1 21 21 I
2013-07-09 12:50:45 1 11 21 21 N
2013-07-09 12:50:44 11 1 21 21 I
2013-07-09 12:43:28 1 11 21 21 N
2013-07-09 12:43:27 11 1 21 21 I
2013-07-09 12:24:44 1 11 21 21 N
2013-07-09 12:24:43 11 1 21 21 I
2013-07-09 12:18:56 1 11 21 21 N
2013-07-09 12:18:55 11 1 21 21 I
2013-07-09 12:12:42 1 11 21 21 N
2013-07-09 12:12:41 11 1 21 21 I
2013-07-09 12:00:49 1 11 21 21 N
2013-07-09 12:00:48 11 1 21 21 I
I: Interference, R: Radar detection, N: Noise exceeded, E: Error threshold exceeded, INV: Invalid Channel, G: Rogue AP Containment, M: Empty Channel, P+: Increase Power, P-: Decrease Power, OFF: Turn off Radio, ON: Turn on Radio
I've got 4 access points out there (approx 15k sq ft spread over 2 levels). According to the dashboard 2 are Ch 11 and 2 are Ch 1 - all have a Noise Floor of -92 to -94 dBm.
The primary complaint seems to be the random disconnects - I'm assuming the clients are getting briefly disconnected when the AP decides to change the channel.
They are currently using AP93, but I'm toying with the idea of replacing 2 of those with AP105 (ongoing project) to see if it will help. I don't believe the main complainer is using a dual-band card, so I don't know if that will be helpful.
I didn't think that a noise floor of -92 to -94 dBm was particularly bad.