After buying both an HP34401A and a Fluke 287 with fresh calibration and data printouts and then buying Ian's excellent PDVS2mini, I did venture down the volt-nut path a small amount. I also sent out my DMMCheck Plus and a few of Doug's references from voltagestandard.com to get everything as up to date as possible and then ran through all my test equipment.
With all of that in mind, it is true that the BM869s is not quite stable enough to leverage the 500,000 count mode for absolute measurements unless you make allowances for temperature drift specifically. But also the 50,000 count mode can drift a bit as well.
If I calibrate the BM869s at 5V at 73F after letting it sit for 30m or so, and then the next morning turn it on at 65F, it is 6 counts low in 50,000 count mode (about 55 counts low in 500,000 count mode), while the Fluke 287 is bang on to the very last digit under the same conditions (50,000 counts). This result from the Brymen is in line with what can be seen in the thread I linked to:
https://www.eevblog.com/forum/testgear/brymen-bm869-and-bm867-batch-calibration-check/Given the price difference and other aspects of those two DMMs, this is completely acceptable as long as you are aware of it.
The 5V calibration point is easy to keep tabs on and if you calibrate it under the current operating conditions, the 500,000 count mode is honestly very good and completely usable in that scenario.
I also did calibrate the 50V setpoint but it wasn't very far off in my case (2 counts maybe at the worst).
So my suggestion is get the BM869s and compliment it with a good 5V reference that you can ship out every so often for calibration.
I found DCmV to be nearly perfect, far better than spec.
AC stuff is tricky. A few seconds with a function generator and you can get pretty much any DMM to read garbage, so I think that is what a scope is for?
As you said, for resistance and capacitance something like the DE-5000 is a good companion.