One site I manage is on a shared hosting server from a major vendor. The site runs a MySQL-backed CMS.
In the past few months, my admin colleague and I have noticed several apparently random occasions when page loading stalls almost immediately after displaying a little bit of generic CMS stuff. No site-specific content at all. We've seen stalls of 15 to 30 minutes. (Inexplicably, no site visitor has ever complained, though there are a lot more of them than us.)
In one case I happened to check the site during a stall and I had time to sign in to the hosting control panel. Suspecting a db issue, I ran phpMyAdmin on the site's db. I got past the preliminaries but the first data query stalled. Eventually, 10 or so minutes later, the problem cleared itself --I certainly didn't fix anything-- and the query returned appropriate data. I checked the site a few seconds later and it was operating normally.
The hosting service tech support has been unresponsive or evasive. In one case, I reported a pure stall and they demanded an error message.
In the most recent stall, my colleague reported this message appeared:
Unable to establish connection to MySQL 2002 : Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (2)
Based on these observations, what is the likely cause of the stalls? A MySQL fault, maybe an overload? Any other possibilities?
What diagnostic logs might be available on the server? What can/should I do to instrument the site?
Does it seem likely that the same underlying problem is being observed, without or with the above-quoted error message? (Maybe the tech staff has recently gotten on the case and added/enabled the message?)
Does this look like something I can take care of or is it a matter of persuading the hosting service to fix?
Are such outages a more-or-less standard "feature" of inexpensive hosting? (The project unfortunately can't afford better.)
Am I missing anything incredibly obvious?