Dan Knauss

I still got two database backup emails and two file change warning emails today. Both sets had the same time stamp, but the file change list was not included.

I’ve never seen that before. The file change notification emails just said “A file (or files) on your site at http://solarismed.com/blog have been changed. Please review the report below to verify changes are not the result of a compromise.”

That was the end of the email.

From another WP site (same host, totally different server environment) that is actively using memcache for query caching, I got two database backup emails that were identical but sent 2 minutes apart.

Yes I am using query caching with W3TC. I’ll try turning that off.

Sorry, that was inaccurate. It was a problem with a theme that was allowing XSS.

If you log uptime with a 24×7 type of service, I will be surprised if you see 99% for a month.

Response time is critical if WP (or any other db-driven app) is a key work environment for 1+ users, if you want visitors to interact with each other (or you), if it’s an ecommerce site, etc.

If you have lightweight pages, if you don’t use a lot of plugins, if you use caching well, and if you are not doing much more than publishing posts now and then, the grid is workable–albeit more costly than faster and cheaper hosts.

If you have much going on in WP, an SQL container on the grid will make it a lot faster (and double the cost) but still not fast to the point that you consistently perceive an instantaneous response in the range of half a second. If you compare with a really fast account, you will start to notice the difference.

Try the latency plugin on vanilla installs on different hosts. I did that along with http response logging with 2 shared hosts, 3 different (gs) accounts, and a moderately trafficked WP site in production on a (dv) earlier this year. The grid’s sql latency stuck out very clearly. You can see this just by hitting WP out of the box with 500-1000 page requests from a load testing script or webwait.com.

The sluggishness of the grid stands out strongly in a way you can feel when you are spending hours working in the UI. Especially with heavier applications like a customized Joomla or Drupal site, on the grid any backend page load usually takes at least 1s+. I’m working with a new install of Joomla+Virtuemart on BlueHost right now, and it’s at the optimal response speed of <1s and often ~.5s. It feels pretty instantaneous, which is usually in the .1s range. After 1s people perceive a delay and their mind wanders.