There are many different levels of Web Hosting.
At Hedgehog, we like to compare hosting to taking care of a car. The lowest level of car care is simply providing you with a parking space so that you have a place to keep your car. This is comparable to the $20-$40/month hosts that provide space for a website on a shared server, but pretty much leave the site untouched when it comes to service.
That basic kind of solution is inexpensive and can adequately meet the needs of static sites without high availability needs.
However, for high-availability, dynamic sites, this kind of solution can be disastrous. Once you have a dynamic site, you usually need a certain level of regular work on your site to keep it up and running properly—the equivalent of routine maintenance, like oil changes on your car. If you never got an oil change for your car, it would run fine for a while, but sooner or later it would blow a head gasket. Guaranteed.
Once you start adding on these kinds of services to your hosting, you enter the world of managed hosting. Though exactly what kinds of services you get will vary a lot from hosting provider to hosting provider.
Many providers out there will do basic support of software and hardware, but assume you have your own administrators on staff to do the fine-tuning. Some will call you to let you know that your site is down, but will not assist in trouble shooting the problem. Others will notify you if a new security patch has come out, but will not install it.
Installation and Setup
This might include building and configure servers and
other appliances, installing and configuring software and applications,
setting up backup systems and monitoring systems.
This might include basic monitoring to make sure the
servers are up, content matches to make sure different parts of the site
are interacting properly, transaction monitoring to make sure users can
perform certain key functions, monitoring memory utilization, disk
space and CPU utilization.
Administration and Support
This might include checking error and maintenance logs,
watching for runaway processes, monitoring table space growth, looking
for file corruption and data consistency, watching background processes,
tuning database files and processes, researching and resolving issues
with load, disk space, errors, and returned mail, as well as watching
for patches and updates, and evaluating and installing them.