What happened during the shift ending 2015-06-16 00:56 (Duration 3h 05m)?
That night one of the central servers went down for about 15-20 minutes. Normally this would only affect the website, while mining servers would be fine. But it looks like we were hashing at a very low hashrate (about 10% of normal) for 10-15 minutes, and after that it took some time for the hashrate to go back up, probably due to miners taking some time to move back from their backup pools. I'm still looking into what the issue was with the main mining server. The website didn't recover right away and was down 30-40 minutes in total. I'm also looking into that.
My apologies for this downtime. A small amount of downtime on the website can happen, but mining should always be available.
On a related note, it would be nice if the shifts page would show my % of the pool for that shift (like on the live stats page) so I could easily confirm that my decrease in Ghps followed the pool's decrease and my % remained relatively constant. Also, although I am aware this was discussed at length previously, I find the "Pay/Hour at 1 TH/s" a useless value. I would be much happier with "Blocks found for which this shift was active" along with my % of the pool. The zeros listed here would be no worse than the zeros listed in the "Pay/Hour" column. The "still eligible for pay" could then be changed to "#[of blocks already found] + still eligible for blocks in the next x shifts".
Maybe I can squeeze in a "your % of the total work" thing just after the number showing your work (rightmost column today).
If space gets tight I could probably show the difficulty in billions instead of the exact number. So "49.7" instead of "49,692,386,354". Or perhaps the difficulty isn't even that interesting in the shifts list?
Yes, a column showing the number of blocks found would be useful. Perhaps showing "0 (still searching)" or "1 (still searching)" etc, for the 10 latest shifts?