I can see you're running the camera in 8-bit mode which is giving you limited dynamic range. So you are frequently bumping up against the saturation level while not getting much improvement on SNR. I saw you tried running with 2x2 binning which is definitely not what you want to do with this problem. Here are some things you can try:
1. Change the driver option to return 16-bit data and change the saturation-ADU setting in PHD2 to 65537.
2. Be sure you're using the latest version of the ZWO camera driver
3. Disable the star-mass checking feature in PHD2 (guider tab of Advanced Settings)
4. Consider reducing the camera gain to improve dynamic range. Note that the gain setting in PHD2 is a percentage of the support range of gains for the camera so you will need to make that translation yourself.
5. Try reducing the guide camera exposure time
6. Don't use an imported dark library, build a new one with your current setup. Changing gain settings also affects the dark library.
7. When you get frequent lost-star events with no obvious explanation, use the 'Diagnostic image logging' feature in PHD2 to capture the guide camera frames. Pull those frames into an imaging app and closely look at the data. It's possible you are getting vibration or other sorts of things that reduce the quality of the available guide stars.
Your guide sequences are very short so I can't get a feel for how often you're losing stars. It seems to vary substantially even while you continue to point to the same sky location. Can you be sure that you're not encountering transparency changes in the night sky? There are several places in the guide logs where you got a huge excursion in RA, something that probably originates outside the mount. Obviously, you are likely to get lost-star events when this happens.
We have lots of users of PHD2 who have this camera so it's clearly possible to get to the right configuration. You can try asking specific questions on other forums to see how others get good results. A quick web search shows up a fair number of discussions in this regard.
Good luck,
Bruce