FastCGI: comm with server means that PHP/FPM has run out of memory

The following apache-error.log, means that the particular web request waited over 600 seconds for a response, and then failed. We have a 10 minutes (600 second) timeout on all Apache requests, to allow your sites to recover gracefully in the event that they were stuck on a long process or overloaded with incoming web requests. 

apache-error staging-12345 [Fri Sep 28 13:35:50 2018] [error] [pid=15367] 208.68.128.91 "https://mysite.com/admin/config/links-checker?clear=all&t=block" "Mozilla/5.0 (Windows NT 10.0; WOW64; rv:61.0) Gecko/20100101 Firefox/61.0" vhost=mysite.prod.acquia-sites.com forwarded_for="xxx.xx.xxx.x" request_id="v-04576240-c322-11e8-a17e-4bede48e238d" hosting_site=mysite FastCGI: comm with server "/var/www/site-fpm/pbgc.dev/php-fpm" aborted: idle timeout (600 sec)

Apache can time out PHP/FPM processes for a variety of reasons, and these are some of the most common ones, along with links to some additional troubleshooting documentation:

• A web request to an external service has failed, and left the process in a 'hung' state. For more information about gracefully handling external calls, visit Handling external website calls.
• Too many incoming web requests are bootstrapping the Drupal application, and there are no available processes to handle the next request. We recommend that you use Varnish to quickly deliver the majority of your content from cache. For more information, visit Using Varnish.
• The PHP memory limit is set too high for all requests, consider setting a relatively low memory limit for normal site traffic, and increasing the memory available for content editors with a Conditional Memory limit: Conditionally increasing memory limits for Drupal and Drush