Our admins maintain each server to prevent or quickly resolve issues that affect all clients on a server. There are many variables that can contribute to single sites or e-mail accounts being inaccessible that do not, unfortunately, 'count' in terms of server downtime, or may cause site speed issues.
If only one client is affected we may not have any knowledge or information about such a delay. Some of these variables include, but are not limited to:
MySQL databases being out of date
-- slow connections from a website application to the database server can result in a site appearing 'down' since most browsers or tracking services only allot a certain number of seconds for a page connection to load before they time out and report errors.
What you can do in this instance is maintain the databases frequently with PHPMyAdmin optimization and repair tools
PHP scripting issues can cause delays or timeouts
-- sometimes PHP scripts that help to run or maintain your site are not optimized for the shared environment.
These can be scripts self-coded or tools included such as themes, plugins, or other site modules. Keeping PHP scripts up to date or ensuring that code for such scripts is optimized for your server's PHP and settings can make a big difference.
Your server is running one of the most recent versions of PHP, which we update regularly. You can also tweak memory, timeout, and other settings locally with a php.ini file in your web directory
Abusive connections via MySQL or Apache
-- if you have search bots or if someone is opening a lot of connections there are limits in the shared hosting environment to how many simultaneous connections can be open at the same time.
You can check this out a bit with some of your Logs tools in cPanel. Whenever someone visits the site it opens at least one connection [sometimes multiple depending on the page content] and if those limits start to fill up from excessive traffic, just on your site alone, we will never be flagged since there are no server issues but your remote tools like Pingdom may report that it was unable to get in and connect.
This last item has a lot of different variables, so I won't be able to cover them all in one article.
Essentially we are mainly able to monitor issues where most if not all of the clients will be affected.
If you or a web developer are able to optimize your site that might help improve overall daily performance. You can see some of the limits for your account via cPanel >> Logs >> Resource Usage. We are running a program called CloudLinux on the server to balance Memory and CPU consumption for all accounts to perform well. (This is NOT run on any service besides cPanel shared, so Clouds/Dedicated servers will not have it available.)
A shared environment is prone to many difficulties as the abusive connections of one user can cause site issues for more users on that same server. Our policies prevent abusive accounts on the server, and like mentioned in previous correspondence we do take action when our monitoring software indicates load above a certain level that should start affecting the users on the accounts. Each account has unique needs.