ERPNext Site down!

I have been running a production ERPNext system for a few months. Debian, on Google cloud VM. as site1.local.

I also have a sandbox site, using port multitenancy, on port 82.

We were editing Item records and suddenly lost connection to site1.local. Restarted bench, restarted the VM with no success. We can access the sandbox site on port 82, but not the main site. We get ERR_CONNECTION_TIMED_OUT

currentsite.txt has site1.local in it
database for site1.local is accessible, using HeidiSQL

What troubleshhoting steps can I undertake?

worker.error.log has “connecting to localhost:11000. Connection refused” a few times
frappe.log shows no errors

What is the size of your VM? Looks like you were out of memory and db crashed.
You will have to reset the VM instance from Google cloud console.

I’m running 1vCPU, 4GB memory. Funny, Mariadb is accessible, using HeidiSQL client, so database is running. Also, sandbox on port 82 is fine as well

Check the monitoring charts on Google console. Is ram usage high at any point?

You probably need to restart your mysql and/or supervisorctl

Like magic, it’s up and running this morning. I upped the VM memory last night to 5GB from 4GB, maybe it took time for Google to actually assign the extra memory to the VM?? Unfortunately, ram usage is not a metric on the Google console.

Thanks for the feedback everyone!

Mike