Attention Required: Your DreamHost VPS needs help!


Every so often I get this message. I have a look on my usage of the VPS and see that my allocation is always 200MB more than my average usage.

This time I see a real large hike in CPU see image below.

I’ve asked support about this several times but they always baffle me by telling me to look into the logs which I never have time to do.

Would be great to hear if anyone has any advice for speed.

I have several WordPress sites on this VPS with low traffic.

I have done all I can to speed up the sites but I’ve never been happy with the speed. Especially of the initial load.

If anyone could help me with advice on this I would be extremely grateful


I have a difficult time trying to decide how to react to your post.

You ask support how to find out about memory spikes and they give you the answer, yet you don’t have time for anything like looking at logs. Do you think they should do this for you? Do you think they have more time than you do? Also you have only given us one tiny snapshot of what’s going on, and haven’t really told us import things, like “the spikes occur or don’t occur with any pattern like the same time everyday.”

Then you change subjects abruptly and want advice about page speed and page loading time. You say you have done all you can to speed up the sites, but don’t tell us what that includes.

Don’t be surprised if people are reluctant to help, while they may be able to offer suggestions you’ve said yourself that you don’t have time to deal with your site’s management. Unlike Staples there is no Easy button.


Basic thoughts:

  1. Upgrade PHP to 5.3 or 5.4 - they use less memory than 5.2
  2. Google PageSpeed is your friend
  3. Caching plugins are your friends

You have … 20 or so sites on that VPS, so they all should be looked at carefully, because they all add up to the total usage.

By the way, it’s a trademark violation to have ‘wordpress’ in your domain name: You should rename that site to wp-whatever and not wordpress-whatever :slight_smile:


Your screencaps look like what you should expect when users visit your site, although the apparent time factor looks odd at first glance (seems to commence precisely on the hour and stop precisely 2 hours later). As Lakerat pointed out, it would be good if you can compare your findings above with those of other days to see if there’s some correlation with respect to time.

For additional speed on VPS you could try opcache like XCache, APC, etc. (W3TC implements XCache well and I notice you already use it on the site I looked at). You can move the sites you require to be super responsive to their own userspace and use XCache via your W3TC plug-in to check improvement in response time.


If you want to find out what is running when your server load peaks, you could use my memlog tool:

As LakeRat stated however you’re going to have to put in some time to find the cause, this includes looking over logs that the tool produces.

From the screenshot, I get the impression that your sites aren’t being visited at all and you just had some visitors that caused this relative peak, or maybe it was a search engine indexing your site, or maybe a job that you run regularly. Whatever it is that is running it does appear rather CPU intensive.


Thank you Ipstenu-DH

  1. All are on PHP 5.3 CGI
  2. I’ve spent hours on PageSpeed and resolved the major issues
  3. I have caching plugins on all the sites – even the ones with negligible traffic

I’m aware of the trademark violation, thank you, – it points you towards another site.

sXi, ottodv, I’ve spent hours looking at the logs all I see is people accessing the site. I don’t really know what I’m looking for, that’s the problem.

I don’t understand why a few visitors or a search engine visiting the site on 15 or so WordPress sites causes a $400 pa VPS to crash.


If you’d have looked at the logs that memlog produces, you’d clearly see which processes used the most resources during those peak times.