Two outages in five days. Neither was a hack, and that is the uncomfortable part

Saturday 15 August, 2:25pm. Mastercard payments stopped working across Australia. Cards, Apple Watch, PayPal. It came back around 4pm, so about ninety minutes, on a Saturday afternoon, which for a lot of businesses is the best ninety minutes of the week. Downdetector logged around 1,700 reports by 3pm, and Australia gets through roughly 44 million card purchases a day, so the number of interrupted transactions was not small.

Wednesday 19 August, just before 10am. Google Drive started throwing 502 errors across Australia and New Zealand. Files would not open, would not edit, would not send. Reports peaked around 1pm and Google confirmed the fix just after 11pm, as SmartCompany reported. That is thirteen hours. A full working day where quotes, contracts, job sheets and client files were simply out of reach.

Two failures, five days apart, hitting the two things almost every business needs to trade: taking money and reaching your files.

Nobody attacked anything

Here is the detail worth sitting with. Neither of these was a cyber attack.

Mastercard put its outage down to “a scheduled system update”, reported by Information Age. Google put its one down to a capacity and routing problem in its regional network, fixed by easing congestion and bringing forward an upgrade it already had planned.

In other words: two of the largest technology companies on earth broke their own systems doing ordinary maintenance. No hackers, no ransom, no villain. Just Tuesday, on a Saturday and a Wednesday.

That matters because it removes the comforting idea that this is someone else’s problem to solve. You cannot buy your way out of this, and there is nothing to defend against. The only useful response is to decide in advance what you do when it happens, because it will happen again, and probably to a tool you have not thought about.

The real failure mode

The outage is not actually the problem. Ninety minutes is survivable. Thirteen hours is annoying but survivable.

The problem is that most business owners discover their dependency list live. At 10am. With a customer standing in front of them, or on the phone, while someone in the back says “it’s not just us, it’s Google”. By the time you work out that your quoting process, your job sheets and your client history all sit behind the same login, you are already having the bad morning.

You want to know that list on a quiet Thursday, not during the outage.

Ninety minutes without payments

Worth being honest about what the fallback actually is now. Only about 5% of Australians use cash regularly, which is why the government had to legislate last year to make essential retailers keep accepting it. “We’ll just take cash” is a plan that assumes your customers are carrying some, and mostly they are not.

The AusPayNet review published shortly before the Mastercard outage called payment systems a critical national service, and warned that failures in software, infrastructure or change processes can scale up fast. That is the polite version of what a florist or a cafe learns the hard way at 2:30 on a Saturday.

The one-page outage plan

This is not an IT project. It is a single page, and an hour of thinking. Five things:

1. List what you cannot trade without. Not every tool. The two or three whose absence stops money coming in: payments, your files, your booking system, your phones. Everything else can wait a day.
2. Get a second payment rail that is not on the same network. If your terminal runs one provider, have a way to take a payment that does not, whether that is a second reader, a payment link you can text, or a card-on-file process your staff know how to run.
3. Keep today and tomorrow offline. Tomorrow’s jobs, addresses, contact numbers and anything you need on site. A printed run sheet in the van costs nothing and works during a thirteen-hour Drive outage. Cloud-first is right; cloud-only is a choice you should make deliberately.
4. Know how you will tell people. If your website, email and files are all with one provider, an outage can take out your ability to explain the outage. A phone with your customers’ numbers in it, or a way to post to your Google listing and socials, covers most of it.
5. Write down who does what. Two lines per scenario. Who checks whether it is you or them, who tells customers, who switches to the fallback. In the moment nobody improvises well.

Do not overreact to this either

The wrong conclusion is to start pulling things out of Google, or to add a second cloud for everything. That is expensive, it doubles your admin, and the replacement will also have outages, because everything does.

Concentration risk is real, and we wrote about the sharper version of it when access to a major AI platform got switched off overnight. But an afternoon offline is a different problem from a vendor turning you off permanently. The proportionate answer to an afternoon is not migration. It is a fallback for the two processes that actually cost you money, and knowing what they are.

The func.digital take

Resilience has a reputation as an enterprise word, and it usually arrives attached to a big quote. It does not have to. For most Melbourne businesses this is genuinely an hour with a notepad, one alternative payment method, and a printed sheet in the van.

What makes it worth the hour is that the alternative is discovering all of it live, in front of a customer, on your busiest afternoon. Two different global companies proved in five days that ordinary maintenance is enough to do that to you.

If you are not sure which parts of your operation stop dead when one login goes down, that dependency map is one of the things we put together in a free digital systems audit. We will find the two or three processes worth a fallback and leave the rest alone. Get in touch, and let’s work it out on a quiet day instead of a loud one.