Our web page load times are killing us. Our CPU usage is peaking out periodically. I checked our visitors log (among other things), and one of our pages is showing up over and over on the log with the same IP address as our web host. This is inconsistent with our Google analytics report. The page in question is archived tech articles that are rarely ever visited.

Anyone know what might be causing this?

I did notice that the previous web developer that scripted our site has made this page clonable. I'm wondering if he did something wrong when he developed the page.

By the way, this archived-tech-articles page has been request 531 times just today. I don't even think we've had that many visitors today!

[trendoman]

Page load depends on overall connection speed and Couch execution time. The latter should be somewhere around 0.5s for an average page. I'll suggest a thorough code inspection as well as .htaccess files. Sent you a PM regarding this.

[KK]

page has been request 531 times just today .. with the same IP address as our web host

Two questions come to my mind -
1. It is not normal for Couchified pages to request other pages directly.

2. Requesting 531 pages in a day is not a number that should cause such peaks if the requests are spaced out (as happens normally on a site). That can happen, though, if those number of requests are made in a burst e.g. within a few minutes.

Prima fascie, things point out to, probably, some kind of script at work on your server.
Needs thorough investigation, as @trendoman suggested.

Requesting 531 pages in a day is not a number that should cause such peaks if the requests are spaced out (as happens normally on a site).

It's weird because there will be a request for the same page four times all at the same time code. In between these bursts of request for the same page will be a normal request such as an image or something, and then there could be 3 - 8 requests again for the archived tech article all with the same time code.

Trendoman offered to take a look at the code in a private message. I'll post if there is anything found in the code that was causing this. Stay tuned!

[trendoman]

Every single template out of 2 that I saw is a mix of php and couch tags. Each language version calls for up to 10 external templates/snippets. I feel there are some JS files thrown in there too, so that counts

Anyway, Couch never experienced anything similar, never been hacked or code injected, afaik. This makes the whole case very interesting.

What version of Couch you have installed?

Edit:
Links on the website are not properly working: I can't switch to Spanish version.

Next, if I select articles, published in 2000, I am presented with an article which eventually shows "Published: July 1st 2007".

Next, links were coded directly without proper support of pretty-urls. Sample:

http://www.mywebsite.com/catalog.php - existing link in code
http://www.mywebsite.com/catalog/ - what comes out after hit on the link above.

This means that there is a redirect for the visitors from first link to second and it affects search engines too, because bots experience many redirects.

That said, I am concerned that the original posted issue is not the only one on that website. There are a lot of things too. I've seen Couch tags in raw php code. Sent you a PM with samples and suggestions. Hope it helps.

[trendoman]

A follow up: CouchCMS version can be seen on the bottom of any template in Admin Panel.

download.png (3.73 KiB) Viewed 2533 times