That does seem quite unusual. I have seen arcs move around on the electrode surface, particularly in new lamps sometimes from the edge and creating a slight kink. My observations seem to relate to the most emissive points on the electrode surface and generally stabilises around the shortest path for the arc. Two arc path sounds even more rare. I have only observed anything like that in tubes I have made with some added iodine. This appears to create two different paths depending on current direction so there is actually a rectifying behaviour separated into two arc paths in the one tube. I have never seen this behaviour in a commercial tube. Also, the kick in the arc shown is large enough to suggest influence from a magnetic field. Taking a second look I wonder if the lower support is common to the top electrode and we are seeing an attraction to the electric field there. Could the second arc be an internal reflection as it's hard to be sure in the picture?

Can't say I've seen it before, not on any of my lamps anyway. Very strange