Controlling crawl speed/delay through dynamic server-code and 503's
-
Lately i'm experiencing performance trouble caused by bot traffic. Although Googlebot is not the worst (it's mainly bingbot and ahrefsbot), they cause heavy server load from time to time. We run a lot of sites on one server, so heavy traffic on one site impacts other site's performance.
Problem is that 1) I want a centrally managed solution for all sites (per site administration takes too much time), which 2) takes into account total server-load in stead of only 1 site's traffic and 3) controls overall bot-traffic in stead of controlling traffic for one bot. IMO user-traffic should always be prioritized higher than bot-traffic.
I tried "Crawl-delay:" in robots.txt, but Googlebot doesn't support that. Although my custom CMS system has a solution to centrally manage Robots.txt for all sites at once, it is read by bots per site and per bot, so it doesn't solve 2) and 3).
I also tried controlling crawl-speed through Google Webmaster Tools, which works, but again it only controls Googlebot (and not other bots) and is administered per site. No solution to all three of my problems.
Now i came up with a custom-coded solution to dynamically serve 503 http status codes to a certain portion of the bot traffic. What traffic-portion for which bots can be dynamically (runtime) calculated from total server load at that certain moment. So if a bot makes too much requests within a certain period (or whatever other coded rule i'll invent), some requests will be answered with a 503 while others will get content and a 200.
Remaining question is: Will dynamically serving 503's have a negative impact on SEO? OK, it will delay indexing speed/latency, but slow server-response-times do in fact have a negative impact on the ranking, which is even worse than indexing-latency.
I'm curious about your expert's opinions...
-
Hi INU,
I always like avoid using things like 503s as a general rule. There is almost certainly a better way to do it.
What about just using Google webmaster tools and Bing webmaster tools? Regarding HREFs it depends how much you rely on that tool. If you don't use it, then I'd more more likely to just block that bot in robots.txt and make sure Google and Bing are controlled using the appropriate tools in the respective webmaster tools.
To answer your specific point about whether or not 503 can hurt rankings. In general no as long as they are only short-term. A 503 like 404s or any other response code is a natural part of the web, however, Google has said in the past that repetitive 503s can be treated as permanent rather than temporary and in some cases can result in the pages being removed from the index.
I hope this helps,
Craig
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
-
Will pillar posts create a duplication content issue, if we un-gate ebook/guides and use exact copy from blogs?
Hi there! With the rise of pillar posts, I have a question on the duplicate content issue it may present. If we are un-gating ebook/guides and using (at times) exact copy from our blog posts, will this harm our SEO efforts? This would go against the goal of our post and is mission-critical to understand before we implement pillar posts for our clients.
White Hat / Black Hat SEO | | Olivia9540 -
Googlebot crawling AJAX website not always uses _escaped_fragment_
Hi, I started to investigate googlebot crawl log of our website, and it appears that there is no 1:1 correlation between a crawled URL with escaped_fragment and without it.
White Hat / Black Hat SEO | | yohayg
My expectation is that each time that google crawls a URL, a minute or so after, it suppose to crawl the same URL using an escaped_fragment For example:
Googlebot crawl log for https://my_web_site/some_slug Results:
Googlebot crawled this URL 17 times in July: http://i.imgur.com/sA141O0.jpg Googlebot crawled this URL additional 3 crawls using the escaped_fragment: http://i.imgur.com/sOQjyPU.jpg Do you have any idea if this behavior is normal? Thanks, Yohay sOQjyPU.jpg sA141O0.jpg0 -
Mobile Redirect - Cloaking/Sneaky?
Question since Google is somewhat vague on what they consider mobile "equivalent" content. This is the hand we're dealt with due to budget, no m.dot, etc, responsive/dynamic is on the roadmap but still a couple quarters away but, for now, here's the situation. We have two sets of content and experiences, one for desktop and one for mobile. The problem is that desktop content does not = mobile content. The layout, user experience, images and copy aren't the same across both versions - they are not dramatically different but not identical. In many cases, no mobile equivalent exists. Dev wants to redirect visitors who find the desktop version in mobile search to the equivalent mobile experience, when it exists, when it doesn't they want to redirect to the mobile homepage - which really isn't a homepage it's an unfiltered view of the content. Yeah we have push state in place for the mobile version etc. My concern is that Google will look at this as cloaking, maybe not in the cases where there's a near equivalent piece of content, but definitely when we're redirecting to the "homepage". Not to mention this isn't a great user experience and will impact conversion/engagement metrics which are likely factors Google's algorithm considers. What's the MOZ Community say about this? Cloaking or Not and Why? Thanks!
White Hat / Black Hat SEO | | Jose_R0 -
Google Penguin w/ Meta Keywords
It's getting really hard filtering through the Penguin articles flying around right now so excuse me if this has been addressed: I know that Google no longer uses the meta keywords as indicators (VERY old news). But I'm just wondering if they are starting to look at them as a bigger spam indicator since Penguin is looking at over-optimization. If yes, has anyone read good article indicating so? The reason I ask is because I have two websites, one is authoritative and the other… not so much. Recently my authoritative website has taken a dip in rankings, a significant dip. The non-authoritative one has increased in rankings… by a lot. Now, the authoritative website pages that use meta-keywords seem to be the ones that are having issues… so it really has me wondering. Both websites compete with each other and are fairly similar in their offerings. I should also mention that the meta-keywords were implemented a long time ago… before I took over the account. Also important to note, I never purchase links and never practice any spammy techniques. I am as white hat as it gets which has me really puzzled as to why one site dropped drastically.
White Hat / Black Hat SEO | | BeTheBoss0 -
For traffic sent by the search engines, how much personalization/customization is allowed on a page if any?
If I want to better target my audience so I would like to be able to address the exact query string coming from the search engine. I'd also like to add relevant sections to the site based in the geo area they live in. Can I customize a small portion of the page to fit my visitors search query and geo area per the IP address? How much can I change a web page to better fit a user and still be within the search engine's guidelines?
White Hat / Black Hat SEO | | Thos0030