Why do you really need this ?
Why not just set up a VPS and deploy the NginxProxyManager to it together with a wireguard tunnel to your home system.
You really dont need 2 proxies...
Or if you want to keep the NPM localy on your home server, then you just setup wireguard on VPS with NAT and port forward to your tunnel.
Most likely low memory, linux systems when there are out of memory calls the OOM killer, which kills the largest processes running on the system. https://unix.stackexchange.com/questions/153585/how-does-the-oom-killer-decide-which-process-to-kill-first
You can probably find evidence of this in the system logs, on most modern systems, this should show you higher prio logs from last boot:
journalctl -p 4 -b
You should see something like this, if it was killed by OOM killer
MESSAGE=Killed process 3029 (Web Content) total-vm:10206696kB, anon-rss:6584572kB, file-rss:0kB, shm em-rss:8732kB
And you should definitely think about extending the memory of the system or reducing the number of containers/they memory footprint.
If you have decent experience in soldering, you can unsolder the coil, unwind at least the several top winds and rewind it again tighter and solder the wire end back to the terminal. If you want to completely replace the coil, you can use even a cheap tester like this https://www.ebay.at/itm/All-in-1-LCR-Component-Tester-TransistorDiode-Capacitance-ESR-Meter-Inducta/402624957700 to measure it (you have to remove the insulation on the end of the dangling wire). Then replace it with a similar coil.
If you have no experience with this or doesnt know anyone who does, just buy a new one as many recommended.
I use gotify for various services - it tells me when new stuff is added to jellyfin, it sends me notifications when uptime kuma detects something is down, if grafana monitoring thresholds are reached, after watchtower updates my docker containers, some homeassistant actions, info that some scripts finished and so on... It has a lots of uses.
I mean its 1 command to store your user crontab and 1 to restore it, so I would not over complicate it...
backup:
crontab -l -u root > /tmp/your-crontab-backup
and restore:
crontab /tmp/your-crontab-backup
My lab is around 65-70W, thats Router + Switch + Server (Ryzen 3600 + 3x 3.5"HDD)