Here are some observations:
1. your calibration steps look about 2x the size i calculated. you might try creating a new profile via the profile wizard and make sure you leave everything at defaults.
2. you don't have any features to improve guide star recognition. I suggest you do a darkframe library, enable star mass detection, and set minimum HFD to at least 2 pixels (avoid hot pixels)
3. on your algorithms, set DEC resist switch aggressiveness to 100%, there's no reason to have it lower. if you have an encoder mount, use lowpass2 as the algorithm
4. RA is your constraint on guiding, and it looks like for this you are dealing with a couple of periodic errors, your primary 385.8 seconds is about 4"
It's not huge, but you can use PPEC algorithm to more effectively tame this. Switch your RA guide algorithm to Predictive PEC (PPEC), use period length value of 385.8 and uncheck auto adjust period. I would bump up both the reactive and predictive weights by 10 each, i think the defaults are a little conservative
5. last thing i'll mention is some very curious things happening with your guide star details. normally guide star SNR and mass are a relatively flat, like this example (this is not your data, and ignore the RA/DEC. pay attention to the yellow line at the top):
However, your star SNR is relatively low and exhibits a fade/jump behavior
Combined with some of your other settings above, it seems like it could be locking on to clumps of bad pixels or other things. I'm not ready to call it on that result since the guiding otherwise looks pretty normal, but it seems to me your guiding challenges are really these occasional 5" jumps.
they also get larger as the night goes on, this is the big picture from your 2h45m last run
I find with these kinds of things it's a chicken and egg problem - is the guidestar causing this, or is some sort of mechanical instability the source and causes the guidestar to jump. you'll need to figure that out. hopefully the suggested improvements above will help sniff that out
I suspect it may be a guidestar problem, since i observe the corresponding RA/DEC jumps actually are steady or decrease as the guidestar jumps in SNR increase
hope that gives you some places to start