Way too many work units are being downloaded to the point that they can't be completed prior to the deadline. What can I do to prevent this?

---

It is due to the max_concurrent tag in your app_config.xml file. Get rid of it.
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=5726&postid=45384#45384

Also you can set "Max # of jobs for this project" in "Preferences for this project" on "Your account" page
The setting there has a max of up to 8 work units per computer
This also stops " max concurrent" being a pain ,
It is a known bug in boinc and as you now know it generates a work unit bomb, DONT USE IT.
You can also set the "Max # of CPUs for this project" per work unit
So with a 20 core cpu and if you set it for 8 you should get two work units using 16 cores total
Or set it for 6 and have 3 work units using 6 cores each, total 18 etc.
The rest is `Do the math` :-)

You need to check your computers. Work through them one by one. Were you trying to run the legacy apps or camb_boinc2docker? Some of your computers are low on memory to be running camb_boinc2docker aka VirtualBox tasks. Those tasks take 2 Gb per task.
I would set Max#CPUS to 1 or 2 in your case and max # of jobs to 1 in project preferences. One of your computers is just giving errors and isn't starting the VMs correctly.

Oops, did`nt see that bit.
"Max cpu`s 4" is as far as it can go
I have got Q9450 and Q9550 still crunching boinc here :-)