Yes, doing serveral Let's Encrypt updates at the same time will for sure break it, because the cert file names change while the web server reconfiguration is not completed.
I think I've had a similar issue before, and the simple solution was to run
# httpdmng --reconfigure-all
This takes all the current configuration values from psa database and corrects the web server configuration files. Please also see
https://github.com/plesk/letsencrypt-plesk/issues/93 and scroll down to the comment where we provide the
cert_emergency_response.zip script. It can rename certificate files on-the-fly to the values given in the web server configuration files. Read comment in the thread and in the file before using this script! It does not provide a final solution, it is only a fast correction method in case you really urgently need to bring the web server back up and cannot wait on reconfiguration of the configuration files.
In order to make a cert work for a "www." subdomain, the checkbox "include www" needs to be checked when the cert is generated. Look into the certificate if it was generated WITH the www part:
# openssl x509 -in CERTIFICATEFILE -text | grep DNS:
If the www part is there, it can be used for the www-subdomain. If it is missing, reissue the certificate WITH www subdomain. Certificates can be found in /usr/local/psa/var/certificates .