These are just general thoughts---I will not necessarily follow up on any of them.
Metadata is baked into file names (and the exif fields), so we can semiautomatically "audit" the files to make sure every capture proceeded as scheduled. However, it would be preferable to check that captures occurred as scheduled right away, so that we could correct problems sooner rather than later. One would probably want to use python rather than bash scripting, and perhaps also sqlite.
rsync
seems pretty solid, so I am less worried about failed transfers causing loss of data, but I would like to record stats of which file transfers worked the first time, second time, etc., to have a better baseline for wifi uptime. (This would be useful for distinguishing wifi problems from crashes.)
Ganglia logs all sorts of stuff, but I do not know how to query it. I also do not have a good sense about when I should start worrying about averaging of old data in the round-robin databases.
As fair as I can tell, Ganglia does not log anything equivalent the output of cat /proc/net/wireless
("Wifi link quality", "Wifi signal level" [in dBm], and "Wifi noise"), nor the output of iwconfig wlan0
(which includes 'Bit Rate' in Mb/s and 'Tx-Power' in dBm).
One can get notifications from Ansible (e.g. via slack): https://github.com/maliagehan/gehan-bramble/blob/master/playbooks/take-pictures.yml