I can see the resynchronisation event and can confirm that the circuit was a little bit unstable just before the re-train occurred.
Unfortunately the uploads to MDWS ceased after 2323 hours (2 August) and resumed at 1007 hours (3 August) so there is a big gap in the data . . .
Looking at the two plots that tell something about the physical infrastructure of the circuit -- The Hlog plot confirms that there is no obvious wiring defect and that although the line length is tending to longer rather than shorter you have adequate signals in the DS1, DS2, DS3, US0 & US1 bands. The US2 band is, essentially, absent. The QLN plot tells me that the circuit is fairly well AC-balanced and that the cabinet is of a significant distance from the telephone exchange with which it is associated. The is minor RF ingress; some (at the lower frequency) are undoubtedly broadcast signals.
For the current behaviour of the xDSL link (across the physical circuit) I would prefer to leave it for others to analyse.