Our site is slow..
-
We have noticed that our site is much slower than many of our competitors and expect that this is affecting our organic ranking.
We are on dedicated UK Fast server with
[20/08/2013 16:44:35] Tony Jackson: Operating System: CentOS 6 64-bit
CPU: 1x Intel Xeon E5-2620 @ 2.0GHz
Memory: 8x 8GB DDR3 RDIMM
Hard Drive: 4x 128GB SSD RAID10
[20/08/2013 16:46:28] Tony Jackson: Nginx, percona, tomcat/solr, Magento CE, cat: 25200 productsI could really do with your thoughts on this page for example
http://www.pretavoir.co.uk/sunglasses/ray-ban-sunglasses.html
-
I am not a huge fan of APC at present (PHP PECL library, not the power battery company). Maybe nginx won't segfault like Apache has been for us.
As for the hosting, what I mean is to use a load balancer with multiple servers. Will you take some performance hit in getting DB data? Probably, but very little. But you're getting far worse now with Magento so it doesn't matter how fast your machine is (like I said, it's a beast so if it's not sitting at 0.01 almost all the time you've got major issues). Remember that this impressive machine is also a single point of failure. I had an impressive box running our DB once (multiple webservers) and the RAID controller crapped out in the middle of the day. We got back up under the two hour SLA but it cost us a lot of sales.
Now we run two webservers, a multi-AZ DB (automatic failover to secondary if master dies) and all behind a load balancer in AWS. I won't pretend this isn't without any issues (we wound up in an odd instance where the DB failover didn't work as expected) but it's a lot more redundant and scalable. I bet our setup is cheaper than your behemoth, too. Remember, you still have to chew through all the code each and every time the page is loaded (your problem is processing the individual thread, not horsepower). Splitting your site amongst multiple webservers means you have scalability as well.
-
Highland, Thank you for the response.
We are using memcached locally as it seems to give best results, having tried a combination of APC and memcached.
Would moving our database to a different server speed up the time to first byte(ttfp) as currently we sit with very little load on cpu? What would be the benefit?
Where you mention 'a cloud based solution' are you referring to using Cloudflare to sit in front of our server to act as a cdn and cache, or to move to a complete cloud based solution? I assume it is their Business package you refer to which gives the ddos and other protection.
Looking at Googles pagespeed tools, out ttfp seems to be the hold-up. which then just leaves the off site code (Snapenage, Comodo, etc) bringing up the last elements.
-
Thanks for this response.
We had been looking at Cloudflare and also at rearranging our server somewhat. Will let you know.
-
Magento is built on top of Zend Framework 1, a notoriously slow PHP framework. It also does some things that slow it down. Some thoughts...
- Be sure your use the Magento caching system.
- Move your database off to another machine. 25k products is a LOT
- Your hosting setup is monolithic. While your box is a beast of a machine (12 threads, 64GB RAM, 4x 128GB SSD) it's still susceptible to slow-down. I bet 50 concurrent users could bring it to its knees. While CDNs can help offload some of this, I recommend the beast with a cloud based solution. I bet you could get 2 or 3 servers plus a DB server for less than what you're paying for this one beast. Then put your machines behind a load balancer. So you would have multiple web servers serving your site up, a load balancer routing traffic to the fastest one and a common DB powering it.
- Switch to Cloudflare for your DNS. They will cache pages for you and they do it for free. They also add a layer of security.
- Use WMT to benchmark your site speed. It tells you the load time in a handy graph. Play around with it and see what happens.
-
We are using Magento on Nginxwth caching enabled with memcache
-
Generating the page seems to take very long. What shopping cart software are you using? Can it cache pages?
-
Thanks for your reply.
This page is a better example of our problem
http://www.pretavoir.co.uk/sunglasses/ray-ban-sunglasses.html
-
Your website isn't that slow from London but there lots of things you can do to improve speed:
1. Use a CDN like NetDNA to host CSS, Javascript and images. By using a CDN, you will also increase the number of simultaneous (parallel) downloads since these files will be hosted on another (sub)domain.
2. You have lots of Javascript files. Try combining them into one file.
3. Add a cache header, this will instruct web browsers to cache your images, css and Javascript files. Simply add this to your .htaccess file:
AddType image/x-icon .ico
<ifmodule mod_headers.c=""># YEAR
<filesmatch ".(ico|gif|jpg|jpeg|png|flv|pdf)$"="">Header set Cache-Control "max-age=29030400"</filesmatch>WEEK
<filesmatch ".(js|css|swf)$"="">Header set Cache-Control "max-age=604800"</filesmatch>
24 HOURS
<filesmatch ".(html|htm|txt|php)$"="">Header set Cache-Control "max-age=86400"</filesmatch></ifmodule>
======================================================
... and sorry for my english.
Hope that helps!
-
Yes had looked at GT Metrix though not paid much attention to YSlow. Will take a look now.
However it may be of more interest to look at the speed of this page rather than homepage;
http://www.pretavoir.co.uk/sunglasses/ray-ban-sunglasses.html
-
Hi Sean,
Your system looks pretty decent... I guess it depends on your traffic load though.
Have you checked out your results on GTMetrix.com?
You scored well in the PageSpeed testing, but not in YSlow. I'd take a look and see if you could leverage some of the suggestions there.
Hope this helps.
Mike
Browse Questions
Explore more categories
-
Moz Tools
Chat with the community about the Moz tools.
-
SEO Tactics
Discuss the SEO process with fellow marketers
-
Community
Discuss industry events, jobs, and news!
-
Digital Marketing
Chat about tactics outside of SEO
-
Research & Trends
Dive into research and trends in the search industry.
-
Support
Connect on product support and feature requests.
Related Questions
-
I need Ecommerce Site who have implemented J-Son Schema for Blog
Hi All, Just I need few sites who have implemented J-son schema on there blog. Looking for ecommerce blog. Hope you guys share with me good site. Thanks in advance.
Competitive Research | | amu1232 -
2 sites with low DA and PA and unoptimised <title>outranking everyone for competitive search term - How ?</title>
Im doing some keyword research for a new client, Ive found 2 anomalies for competitive search terms where sites with unoptimised <title>tags, poor backlinks and other stats, (compared to the other sites listed in the SERP) are out ranking everyone and claiming the #1 spot</p> <p> The 2 search terms are :</p> <ul> <li dir="ltr"> <p dir="ltr"><a href="https://www.google.co.uk/search?q=interior+designers+london&oq=interior+designers+london&aqs=chrome..69i57j69i65j69i60j69i65l2j69i60.2784j0j1&sourceid=chrome&ie=UTF-8">Interior designers London</a>- #1 new-id.co.uk - <a href="https://mza.seotoolninja.com/researchtools/ose/links?site=new-id.co.uk&filter=&source=external&target=page&group=0&page=1&sort=page_authority&anchor_id=&anchor_type=&anchor_text=&from_site=">OSE info</a></p> </li> <li dir="ltr"> <p dir="ltr"><a href="https://www.google.co.uk/search?q=interior+designers+london&oq=interior+designers+london&aqs=chrome..69i57j69i65j69i60j69i65l2j69i60.2784j0j1&sourceid=chrome&ie=UTF-8#q=interior+designer+london">Interior designer London</a>- #1 nstudio.co.uk - <a href="https://mza.seotoolninja.com/researchtools/ose/links?site=http%3A%2F%2Fnstudio.com%2F&filter=&source=external&target=page&group=0&page=1&sort=page_authority&anchor_id=&anchor_type=&anchor_text=&from_site=">OSE info</a></p> </li> </ul> <p> (interestingly they hardly rank at all for any modification of those search terms)</p> <p> The only thing i could think of is that they both had some really good links using that exact anchor text, but ive been checking OSE for the past month both in the backlinks and the just discovered but couldnt see anything like that.</p> <p>I conducted the SERP checking from a laptop with a broadband connection based in London, although im not very local to either of these locations im about an 8th of the cities distance away from one of those location and about half a cities distance away from the other. So i dont think its a local skew thats causing the SERP rankings.</p> <p>For both sites i can see they had old domains which now redirect to the them :</p> <ul> <li dir="ltr"> <p dir="ltr"><span data-mce-mark="1">New-id.co.uk = </span><span data-mce-mark="1"><a href="http://newidinteriors.com/">newidinteriors.com</a> - <a href="https://mza.seotoolninja.com/researchtools/ose/links?site=http%3A%2F%2Fnewidinteriors.com&no_redirects=1&filter=&source=external&target=page&group=0&page=1&sort=page_authority&anchor_id=&anchor_type=&anchor_text=&from_site=">OSE info</a> - PA1, DA1</span></p> </li> <li dir="ltr"> <p dir="ltr"><span data-mce-mark="1">Nstudio.co.uk = intarya.com - <a href="https://mza.seotoolninja.com/researchtools/ose/links?site=www.intarya.com&filter=&source=external&target=page&group=0&page=1&sort=page_authority&anchor_id=&anchor_type=&anchor_text=&from_site=">OSE info</a> - Da 32, PA43</span></p> </li> </ul> <p> Although the intarya domain was quite strong with good backlinks (but nothing above par comapred to the other sites the nstudio.co.uk site is now outranking, the anchor text was mostly natural with only some keyword anchor text from PA1 pages. </p> <p>The <a href="http://newidinteriors.com/">newidinteriors.com</a> didnt seem to have a backlink profile, both PA1 and DA1.</p> <p> Ive taken screenshots of the SERP pages below, but feel free to dig further</p> <p>Any idea what's going on here ?</p> <a download="HWkjmDf.png" class="imported-anchor-tag" href="http://i.imgur.com/HWkjmDf.png" target="_blank">HWkjmDf.png</a> <a download="A70jXse.png" class="imported-anchor-tag" href="http://i.imgur.com/A70jXse.png" target="_blank">A70jXse.png</a></title>
Competitive Research | | mike8780 -
How can stale, poor quality sites rank above fresh, high quality sites?
Dear Community, How is it that a Google search for temperature sensors shows the second result as site **A **(http://goo.gl/zRS1L) in a higher positionthan site B (http://goo.gl/3Xy2O) which is on the 2nd page of the SERPs. Site A blatantly does not follow Google's webmaster guidelines and recommendations for design, content or code. Site B scores an 'A' for on page analysis using SEOmoz, it also has a much higher number of root domains in and Moz rank, we have checked that site B has only quality inbound links, and anchor text is natural, it follows Google's guidelines for content and code diligently, and it is on an older domain. We are satisfied that site B has not been penalised by Google, e.g Penguin algorithm updates. We have also done a website verification with Google. Site B has fresher content compared to site A. Site B also has a sitemap yet site A doesn't. We're satisfied that we can rule out the following Algorithm Factors: Penguin algorithm updates Google Notified penalties Dodgy Inbound links Dodgy Anchor text Website verification with Google Fresh vs. static content Sitemap present Server Speed Server location Domain age and about 170 other algorithm factors Does anyone have any ideas as to why site A is ranking higher than site B in Google.co.uk? Thanks in advance. CD.
Competitive Research | | chichesterdesign0 -
Keyword Search - another site bypassed me and they are not ranked well...
I was trying to optimize the keywords "music for kids" . This is a no brainer because my URL is "musicforkids" My On-Page Report Card for "music for kids" gives my site a "A". The on page report card for the other site gets an "F". But the other site comes up number ONE in the organic Google search and my site comes up number three. The two sites above me receive a crappy on page Keyword report for the key words "music for kids" but they still come up above me for that search term. What's the deal?
Competitive Research | | musicforkids1 -
Manta inbound link in OSE - competitor site yes, my site no, why?
Our site has had a link back to our site from Manta for a long time, but it's not appearing in our OSE results. One of our competitors similarly has an inbound link from Manta and that link DOES appear in their OSE results. I'm not certain whether that matters for us. Does it? But in any event, why would the one link appear but not the other? Thanks! Tim
Competitive Research | | tcolling0 -
Free SEO tools appearing in SERPS for a site
Hi there, I've just taken over a site and it doesn't have a huge amount of shelfspace on the SERPs for it's brand name. Looking at the results that are there are a lot of free SEO tools that are coming up. They haven't had an SEO before so it's not coming from their side. There's such a lot of them I was wondering if anyone had come across this before - is it simply a competitor looking at the seo, could it be one of the seo analytics packages using free sites or is there anything else to be aware of. I can start knocking the sites down the SERPs but just wondering if anyone has a good solution for this. thanks Shiv
Competitive Research | | Shivvyt0 -
Is there a way to establish how old a link to a site is?
I wonder if anyone has any insight on the above question.
Competitive Research | | InTouchMarketing560 -
Search shows site links, but not hyperlinked?
2 of our competitors on Google have site links show in their results which brings you to that specific page. In our listing we show up with the same results a few spots behind them with our site links visible showing very similar content, but are not actually hyper linked. Is there a way to fix the issue and let people actually be able to click on our site links instead of just reading them?
Competitive Research | | sknott0