There have been multiple revisions since that build and a lot of fixes with very little documentation about the details of what they fixed.
I would recommend updating to the latest build 3116.000 and also add a button on the HMI somewhere not on a main page that will reboot the controller.
I have not used the DA300, but have had connectivity issues with 2 of my CR3000 HMIs. I can clear up the communications with a reboot and can do it remotely by assigned a button the function CommitAndReset(). I think this has happened even after updating them to the latest build.
Also, I always have flag tags for each device the unit is set up to communicate with that monitor IsDeviceOnline(x) where x is the device number or name. With Modbus, you have to enable the Ping Holding Register for this to work. By default the Ping Holding Register is enabled and set to holding register 1. I always add an alarm (Active Off) to that tag and most of my Crimson apps have Mail Manager set up to notify the operator when selected alarms occur, so rather than having to constantly check on it, it can reach out to you or at least show an alarm on your banner or alarm lists to tell you the comms are broken.
Of course if the Ethernet port of the hardware is completely hosed up when this happens, it can't send messages either. In my cases, it appears to just lose comms to the PLC (A/B). I am assuming that the remote view of the built in web server still works...in my CR3000 that have had issues, that part still functions fine.
I have had log files stop recording due to problems with SD cards. I have also had to tweak some settings in the Web Server for Crimson 3.1 in order to get it to play well with VPN routers and slow internet connections and render correctly in Chrome. I ended up turning compression off for best results (Web Server > Advanced tab > Compression section > Compress Reply: Never)