logoalt Hacker News

Revolut down as users report app not working

53 points • by baal80spam • today at 9:54 AM • 25 comments • view on HN

Comments

monneyboi • today at 10:36 AM

I just learned the embarassing way that my ability to pay for groceries depends on the livelyness of europe-west2.

➕ show 3 replies
GettingOld • today at 11:21 AM

Google cloud is having issues in London impacting HSBC, Lloyds & Bank Of Scotland in UK - May impact your ability to pay with a card

https://www.theregister.com/paas-and-iaas/2026/10/20/google-...

https://www.bbc.co.uk/news/live/c58jd2kjp93xt

Ecstatify • today at 11:15 AM

Revolut needs a way to proactively notify users when their cards aren’t working. It’s so frustrating having to go to Reddit just to find out that Revolut is down.

robtaylor • today at 10:33 AM

Its a bigger thing https://status.cloud.google.com

➕ show 1 reply
afhammad • today at 10:37 AM

Good opportunity to share some competitive resilience: https://monzo.com/blog/tolerating-full-cloud-outages-with-mo...

BoumTAC • today at 10:11 AM

Can we still pay ? I have some shooping to do during my lunch break and I don't want to feel stupid without a way to pay at the checkout.

➕ show 4 replies
TidyMarble92 • today at 11:22 AM

[flagged]

vertias3u • today at 11:06 AM

[dead]

rvz • today at 10:17 AM

Once again, blockchains that use USDC continue to run without any serious outages for years. I just paid for an item with USDC (using Phantom Wallet) right now and there are many other examples of this.

No need to wait for a bank and it doesn't matter the blockchain, as long as it supports USDC and you have a crypto wallet, the payment goes through.

Stripe knows this already and crypto payments will be here for a very very long time as they already are today.

➕ show 2 replies
pllbnk • today at 10:43 AM

Not saying it's a vibe-coding issue. But I just imagine a scenario where a global service experiences significant disruption due to a vibe-coded release. No person understands how it works or is able to figure out in less than a week because no human has ever glanced at the code for a couple years, so the decision is taken to unleash LLM agents directly on production to fix the issue. Of course, one of these agents has to do some database migration but there is a breaking change, so they launch a sub-agent to deal with that; the sub-agent doesn't have a full context, so it decides that the easiest way would be to drop the database and so it does. The latest _successful_ backup took place a month ago.

$100B value down the drain.

➕ show 1 reply