Load Shedding-Proof Websites: Why Your Site Shouldn't Go Down When Eskom Does

Most website hosting advice is written by people who’ve never had to think about rolling blackouts. In South Africa, load shedding is a genuinely normal part of running a business, and it’s worth being specific about one thing upfront: a properly hosted website almost never actually goes down because of it. The confusion is worth clearing up directly, because a lot of business owners assume otherwise.

Why your website usually isn’t affected

If your site is hosted properly, on real cloud infrastructure like AWS rather than a server sitting in your own office, it lives in a data centre with its own power redundancy, completely separate from whatever’s happening at your physical premises. Your shop or office losing power doesn’t take your website with it, the same way your website going down doesn’t stop your till from working. They’re separate systems, and good hosting keeps them that way.

Where load shedding actually does hit a business online

• You personally can’t access the internet during an outage to manage orders, respond to enquiries, or update anything, even though the site itself is fine.
• Customers trying to reach you by phone or in person can’t, even if your website and WhatsApp are working perfectly.
• If you’re running a legacy setup with a server physically on your own premises (rarer now, but it still happens with older custom systems), that server does go down with your power.
• Payment processing can be affected if a customer’s own connectivity drops mid-transaction, which is a customer-side issue but still shows up as your problem in the moment.

What "proper hosting" actually means here

This is the same distinction covered in the AWS versus shared hosting piece on this site, but it matters even more in a load-shedding context specifically. Budget shared hosting is often run on infrastructure with far less redundancy behind it than a major cloud provider, and "cheap" hosting sometimes means exactly the kind of single point of failure that a stage 4 outage exposes at the worst possible time.

What actually keeps a business resilient

• Cloud hosting on infrastructure with real redundancy, not a single server with no failover.
• A mobile-friendly, low-data site, since customers browsing during an outage are often on mobile data with a weaker signal, not fibre.
• WhatsApp Business set up properly, so customers have a way to reach you that works even if you personally can’t get to a laptop.
• A backup power solution for your own premises, if the business genuinely can’t function offline during an outage, which is a business continuity question, not strictly a website one.
• Automated order and enquiry notifications that reach your phone, not just an inbox you can only check once power’s back.

The businesses that get this wrong

The pattern I see most often is a business that’s invested properly in their website but never actually tested what happens end to end during an outage: can a customer place an order, can that order reach someone, can that someone actually fulfil it, all while the shop itself has no power. The website holding up is only useful if the rest of the chain does too.

Worth actually checking, not assuming

If you’ve never deliberately tested your own site and order flow during a real outage, that’s worth doing once, deliberately, rather than finding out during stage 6 on your busiest week. An infrastructure audit covers exactly this: where the site is actually hosted, how it’s configured, and where the real single points of failure sit, load shedding included.

Related