Unfortunately all the logfiles will have been lost as they are held in RAM and a reboot starts with a clean slate. In future it is well worth checking or at least copying the logfiles, you can download the emoncms.log and emonhub.log via the emoncms admin and config pages or you can ssh in and manually copy the logfiles to the data partition.
(1) It’s a bit early to say, we would need more info, possibly if it occurs again you could get the logs. Can you tell us more about where the data is coming from, is it just the standard emonPi inputs or something else? What is the update interval? What interval PHPfina feed is it?
(2) As above, the logs are held in RAM to reduce disk writes to the microSD card, the OS is predominantly read-only and only the data partition is RW.
(3) Looking at the csv output in the Graphs page might help, or try zooming in on the graph (see Solar App - Graph problem)
I have previously seen this behavior in 3 instances, when either the server time or browser times are incorrect (can’t recal which), when the feed interval is less than the input interval (as per the linked post) or on emoncms.org when the input queues get a bit backed up and until we know other wise we should also bear in mind there is a chance this is caused by the way you have your mySQL on a remote server (Couple of questions about the SQL and Backups)
Because there are many layers between the data being generated and sent or even received at emoncms and the final graphing, you shouldn’t necessarily assume an error on a final graph is due to data not being received, the simplest checks are to look at the input and feeds pages to see when the suspect input and feed updated last. Then check/copy the logs, rebooting is a sure fire way to ensure some data loss, not only is there always up to 60secs of data held in emoncms to write to disk at anytime, there is also the reboot time and if there were any recoverable issues that were causing data to not be saved yet it could still be in redis and potentially save able, I’ve had disk space max out before now, preventing data from being saved and when I deleted some other stuff, the feeds just caught up. It may well turn out that you do need to reboot, but those additional couple of mins double checking things and getting the logs will not add much to your woes if you have already lost hours of data.