Every year a few affiliate marketers and small e-commerce operators have a really bad Black Friday. Not because the traffic didn't show up, but because their hosting fell over the moment it did. The sale they spent six months preparing for evaporates in the first hour because the site is throwing 503s at the actual customers.
This isn't really an article about Black Friday specifically. The same logic applies to any traffic event you're expecting: a TV mention, a launch, a Reddit post that goes viral, a guest post landing on a high-traffic site. But Black Friday is the big predictable one, and there's still time to do something about it.
What Actually Breaks Under Traffic
In my experience watching monitors during traffic spikes, sites tend to break in a few specific ways:
Database connection limits. WordPress sites running cheap shared hosting often have a hard cap on database connections. Hit it and the site starts throwing white-screen errors or generic "database connection failed" messages. Catastrophic for conversions.
PHP-FPM workers exhausted. Your server has a fixed number of PHP processes that can run at once. Once they're all busy, new requests queue up. Once the queue is full, requests get dropped. Visitors see timeouts.
Response times balloon. Even if nothing actually breaks, response times can go from 200ms to 8 seconds. Conversion rate craters. Most people don't wait.
Origin server gets DDoSed by your own success. Without a CDN, every request hits your origin. Traffic spikes that would barely register with caching can knock the origin offline.
Affiliate tracking links get rate limited. Some networks rate-limit individual affiliates. A traffic surge can trip the limit and your clicks stop redirecting properly.
What to Check Before Traffic Hits
A week or two before your traffic event, do these checks. None take long.
Response times under load. Your site might be fast right now because nobody's on it. Hit it with twenty parallel requests using something like siege or a load testing service and see what happens. If response times go from 300ms to 4 seconds with twenty users, you have a problem at fifty.
Hosting plan headroom. Look at your hosting dashboard for current resource usage. If you're at 60% of your CPU or memory limit on a normal day, you have no headroom for a 5x traffic event. Upgrade now or accept that things will break.
SSL certificate expiry. Black Friday is the worst possible day for your SSL to expire. Browsers block the site entirely. Check yours expires nowhere near peak traffic dates and that auto-renewal is configured.
CDN configuration. If you're not on Cloudflare or similar, get on it. Even the free tier saves you from most origin overload scenarios. Make sure caching is actually working by checking response headers for cache hits.
Redirect chains. Affiliate offers especially. Test every redirect destination still works and lands on the right page. Networks tend to pull or modify offers at exactly the wrong moments.
Email deliverability. If you're sending campaign emails on the day, do a test send through your own email and check it lands in inbox not spam. Last minute domain reputation changes happen.
Set Up Alerts That Actually Wake You Up
Most affiliates I know have alerts set up but they all go to email. On a busy Black Friday morning your inbox is going to be a disaster zone. The downtime alert ends up between two newsletters and a customer service email and you don't see it for an hour.
Set up Telegram or push notifications for downtime specifically during your traffic event. Drop the threshold to one minute checks for the duration. After the event, switch back to whatever you normally run.
WebMon's slow response alerts are particularly useful here. You don't just want to know when the site is fully down. You want to know when it's slow enough that conversions are dropping but the site is still technically up. Set the threshold somewhere around 3000ms and you'll catch performance degradation before it becomes total failure.
When It Goes Wrong Anyway
Despite preparation, something usually breaks. Have a plan:
- Know which page absolutely must work (your highest-converting offer or your main funnel). Monitor that one most closely.
- Have your hosting support's contact details somewhere you can find them at 4am.
- Have a backup landing page ready to redirect traffic to if the primary fails. This sounds like overkill until the day you need it.
- Don't panic and start changing things. Most outages resolve themselves in a few minutes if you don't make them worse.
After the Event
Whatever happened, look at your monitoring data afterwards. Where were the response time spikes? Did anything fail completely? How long did recovery take? This is the data that tells you what to upgrade or change before next year's event.
The cheapest insurance for a high-traffic day is to look at last year's monitoring data and fix whatever broke. Most people don't do this. Don't be most people.
Set up traffic event monitoring from your WebMon dashboard. Free plan covers basic uptime; affLIFT plan adds slow response alerts and 5-minute checks.