In the old Enhance, Docker would monitor services and restart them on failures/crashes. To restore that functionality you can create systemd overrides. They're pretty simple, and this is normal config editing for systemd, nothing crazy. Here's how you can add an override for mysql that will automatically restart mysql after a crash:
First open the override file:
sudo systemctl edit mysql.service
Next, add this to your overrides file:
[Service]
Restart=on-failure
RestartSec=10
Now reload systemd:
sudo systemctl daemon-reload
Last, restart MySQL:
sudo systemctl restart mysql.service
There should now be a override.conf file in /etc/systemd/system/mysql.service.d/, that file should just contain the 3 lines of customization you made.
You can do this on all your DB servers, so they'll all restart MySQL if it crashes, help you prevent extended downtime. I set restart delay to 10 seconds, so if there's a critical runaway looping error - there's time between restarts where I can stop services or other troubleshooting. You can increase this to more time if you like, whatever balance between uptime and recovery is best for you.
Important Note
I set my override to "Restart=always". The difference being that systemd will restart MySQL no matter why it is shutdown (other than using systemctl command specifically). This is important for me because for the first time ever on Enhance we got a downtime yesterday on a server where the DB was offline. MySQL log shows "user signal" initiated shutdown (wtf). Having "Restart=always" might have caught that and restarted mysql, whereas "Restart=on-failure" would not done anything... So choose how important the DB server is in your environment and add the config as needed, for me it's "Restart=always" with a 10 second delay.
