Has your site ever went down due to so much traffic?

[Cedric]

Have you ever experienced the thrill of watching your website's traffic skyrocket, only to then experience the panic of watching it crash and burn? It's a problem that many webmasters dream of having, but it can also be a nightmare if you're not prepared.

So, we want to hear from you - have you ever had your website go down due to a sudden surge in traffic? Maybe a post went viral or a product launch exceeded your wildest expectations. Did your hosting provider fail to keep up with the influx of visitors, or was there another issue at play?

If you've been through this before, we'd love to hear about how you handled it. Did you have a plan in place, or did you scramble to fix the problem on the fly? What steps did you take to prevent it from happening again in the future?

It's important to monitor your resources and server capacities once in a while and upgrade your hosting package along the way.

And for those who haven't experienced this problem yet, let's talk about what we can do to prepare for it. What measures can we take to ensure that our website stays up and running, even during a surge in traffic? Join the discussion and share your experiences and tips for dealing with high traffic situations!

 

[Arantor]

I've never had a situation where I've had a load scale up so massive I couldn't possibly keep up with it - never had the Slashdot effect, as it were, happen to me.

That said, I tend to healthily over-provision servers just in case. I usually have servers and services provisioned so that there is some headroom in case of traffic spike; normally I can soak a doubling of traffic with zero drama, 3x-4x for short bursts - the platforms I work on, I've spent enough time with to know their performance handling characteristics and know what I can do to scale them and how, I could probably scale up to cope with 5x standard load in an hour, 10x load in a day, as long as these weren't long term bursts of traffic. Would need a bit more time for that.

I was asked last year to take an e-commerce site that could barely cope normally and make it run on Black Friday - if you know load is coming and you have some idea how big it is, planning for it should be fairly straightforward unless the platform is so poorly engineered it can't be fixed. The trick here is knowing what load you take today, some idea of what load is coming and figuring out how much you need to cut out to give you that performance back.

It is also much easier than it used to be to handle such things - with CDNs and things like CloudFlare being more affordable and more available than ever.

Though I will say the worst experience I had was just after launching a revamp of a site, where it was already using multiple servers (big site, millions of hits a month, heavy platform to do the work), the revamp took more resources than the earlier version but the new version's database instance on AWS was under-provisioned (I didn't set it up), to the point it had less than half of the resource it should have had, but I still pushed it to be acceptably performant for existing load (given that it spiked with 'look at the shiny new version') just because I know the platform, I know how and where I can make performance gains and when properly resourced, the thing positively flies.

If in doubt, get someone in who really knows the platform you're working on, and knows how to actually do performance management - but it never hurts to be pre-emptive about it. Not enough sites do this until it's too late.

I don't want to be more prescriptive than that, because it's genuinely hard and scaling up requires trade-offs in most cases so I prefer to do it situationally.

 

[Tracy]

I wish... but with a 4GB/4vCPU/160GB VPS and it being relatively well-tuned, it would take quite a bit of traffic to bring it down.
Especially when you add CloudFlare into the mix.

 

[Arantor]

I wish... but with a 4GB/4vCPU/160GB VPS and it being relatively well-tuned, it would take quite a bit of traffic to bring it down.
Especially when you add CloudFlare into the mix.

The site I mentioned that could barely run normally that I had to fix for Black Friday was WordPress/WooCommerce on a 16GB, 12vCPU VPS. And there wasn't anything wrong with the server, it was coping quite well in spite of the 7000+ queries per page it was doing... sites might not scale as nicely as you might think.

 

[Tracy]

The site I mentioned that could barely run normally that I had to fix for Black Friday was WordPress/WooCommerce on a 16GB, 12vCPU VPS. And there wasn't anything wrong with the server, it was coping quite well in spite of the 7000+ queries per page it was doing... sites might not scale as nicely as you might think.

Depends on many items..... One of that sites issues was WordPress/WooCommerce.
As with ALL things.... you have to have a solid underpinning. Expecting a 1/2 ton truck to be able to carry 12 cubic yards of dirt is not realistic. ALL sites have to work on the underpinning as they grow.

 

[Arantor]

I spent 3 months, made them a new theme, ripped out Elementor, and they were taking money on Black Friday like a champ; the thing never slowed to more than 3 seconds per page load at any time I was watching, and often down to under 20 queries a page when I was done - still using WP/WC!

 

[Tracy]

The pity is one has to jump through as many hoops as you did on a script that is.... to be blunt, not the most efficiently coded.
A server that would bog down under WP/WooCommerce with a set number of visits would not perform as bad with XF. Same way that a site with the same number of visitors performs worse under IPS 4.3 than it did under the same load on XF (ran some load tests on both whenI played with the comparable release versions).

Meanwhile...I continue to get this regularly here on threads.


A refresh does away with it though.

 

[Arantor]

to be blunt, not the most efficiently coded.

Ain't that the truth.

But I come from lands running heavier software than even WP/WC at times - Moodle is a real bear to run at any kind of *actual* scale. (Once did a site specced for 10k concurrent users. We worked out it needed 60 instances on EC2 to cope. By comparison I've seen SMF run 10k concurrent users with 2 comparable instances.)

 

[DigNap15]

One of the problems is that many webhosts do not undesrtand the needs of forums. Eg mySsiteground are obssessed with Wordpress.

 

[Arantor]

Given how much more business they have with WordPress, why *would* they care to specialise in forums?

 

[RankWebSEO]

One time ..... But its due to hosting provider .... hosting was on GoDaddy and when we change hosting then site is running smoothly without any issue till now .... One of our website hosted on Godaddy ... Theme is a bit unique "Needy and poor people" ... So after few months when we add content on regular basis ..... website starting to get hits for many keywords .... I dont think we were receiving much traffics but site go very slow day by day ..... Godaddy support say that need to upgrade package to handle more traffics .... We switch to another hosting and all things are fine now.

 

[Ravenfreak]

I've never had this happen to me before, though I have had a influx of spam bots because I forgot to set registrations to email approval that ended up getting my hosting suspended until I fixed the problem... Thankfully it didn't take long to get things back up and running.

 

[Oblivion Knight]

I've had one or two cheap shared hosts in the past that couldn't handle any volume of traffic with a XenForo installation. I learnt my lesson and decided to invest a bit more in someone who came highly recommended (MattW), and I haven't looked back since.. I'm on one of his shared hosting plans, but if the time came to upgrade I certainly wouldn't be looking elsewhere.

 

[Heatman]

This haven't happened to my website so to speak but I have been on several forums or website where it happened. The first one was a paid to post website called TCP if remember correctly. It got so much traffic that the hosting used couldn't carry all the traffic.

Another one was a job website where thousands of people was trying to register for the job because of quick deadline given. The site crashed because of too much traffic.

 

[Starforce]

I wish that I could say that I am competent enough to have avoided a "crash and burn" moment in my years of self-hosting my communities, but alas, here I am.

Ultimately, it boils down to how efficient your code is, and the resources at your disposal to serve requests. In my case, I didn't dedicate as many resources as I actually needed to the project, and when we had the slightest spike of activity, boom -- Down goes the website.

If you have the resources to spare, it's better to over provision. In the end, you'll give yourself room to grow, but you'll also give your community a fluid experience with minimal disruption.