OK thanks for keeping me in the loop.
Increasing the connections sounds like a good idea even if it did not solve your problem on its own. In fact this whole problem sounds like a matter of scaling your infrastructure, and is a DBA task. If you can point to some error that the plugin produces as you did before, then I can help you with that.
If I were you I would very carefully read the server logs, including the mysql logs, the webserver logs and the wordpress logs around the time when the site start failing.
If you need to exclude some lines from your logs you can use grep -v and pipes to remove some patterns that repeat frequently. Hopefully the real error will stand out and you will be able to know what to do to scale the services.
For example to remove lines that have foo or bar in them I would do the following on my setup:
tail -f /var/log/nginx/access.log /var/log/mysql/{mysql,error}.log /var/www/wordpress/wp-content/debug.log | fgrep -v foo | fgrep -v bar
This would let me observe all the services for any errors in real time but without having the screen cluttered with irrelevant messages. Your log paths may vary.
Keep in mind that the plugin’s frontend hits the webserver (and therefore the database) pretty hard due to the polling feature. One idea would be to disable polling. This would cause the frontend to not update automatically. See Admin → Wallets → Frontend settings → Live polling settings. Set the two polling intervals to 0 (disable) and also disable loading when page becomes visible. That should reduce the load on your server by a lot.
If you do find any other error in the logs, please share it with me.
best of luck!
kind regards
