It would be interesting to check the statistical significance of an observed production fluctuation from what have been previously observed. In short, is the present fluctuation above or under the noise level?
I haven't done that because I have no confidence that the errors in this data are random rather than systematic. However, visually, the recent months look well within the range of fluctuations reported over the last 12-15 months.