My hub ran perfectly fine for 30 days straight on 2.1.5, from Oct. 31st until I updated to 2.1.6 at the beginning of December. Prior to this I could usually go 2-3 weeks between reboots; still unacceptable for a system that I want to simulate occupancy for the length of a nice vacation.
The installation of 2.1.6 somehow knocked the Zigbee radio offline on its initial startup; rather than play with it I restored 2.1.5 and waited for the next firmware. Running on 2.1.7 since early December, the elapsed time between required reboots has not exceeded 2 days. I know I need to reboot when trivial automations randomly fail, error messages begin to appear in the log, and the mean time of virtual switch delays (monitored by Hub Watchdog) doubles... followed shortly by sporadic and increasingly frequent virtual switch max delays exceeding 5 seconds (typically they cluster around .04s). Within few hours of this happening, just about all timed or button-triggered automations begin to suffer noticeable delays or do not work at all. Z-Wave, Zigbee, virtual switches are all affected. Hub is a C3 FWIW; though I doubt the vintage of the hub hardware matters for this scenario. I installed 2.1.8 yesterday. Running well so far, as is usual after every reboot.
I wonder how my hub would run if it were capable of running one of those old backups I have that were less than 10MB in size (my backup file has quadrupled in size over the last year, once increasing by 10 MB overnight for no apparent reason-- when I saw that I did a restore and it has hovered around the same 30MB size since). I probably have added less than a half dozen physical devices over the last year, and probably a dozen new automations (many of them trying to keep track of when HE is going down the tubes and auto-reboot it).
