Revolut Is Down: Why One Point of Failure Costs You
By: Evgeny Padezhnov
One app stops responding, and the card in your wallet stops working with it.
Revolut outages show the same pattern every time. People cannot pay at checkout, cannot move money between accounts and cannot reach support through the app they would normally use. The outage itself is the provider's problem. Having nothing else to fall back on is the customer's problem.
The same failure exists at every scale. A traveler carries one card. A freelancer gets paid into one account. A subscription business runs every charge through one payment processor. Each of them has a single point of failure (SPOF), meaning one component whose failure stops everything that depends on it.
This article covers that dependency at three levels: personal finances, payment flows in a business, and fintech infrastructure. Each level ends with concrete steps.
What a Single Point of Failure Actually Costs
Flexential describes the immediate impact of a SPOF as downtime. Downtime disrupts services, customer service and internal communications. The second-order effects are larger: financial losses, operational disruption and damage to reputation, including lost investor confidence.
For payments, the damage is easier to count. BR-DGE estimates that failed payments cost the global economy hundreds of billions each year. Many providers route all transactions through one system or gateway. When that system fails, the whole payment flow stops.
Recurring revenue is affected too. Zethic cites Recurly (2024) data showing failed payments account for 20–40% of total subscription churn. Most of those customers did not decide to leave. Their payment simply did not go through.
In plain terms: an outage costs more than the hours it lasts. It also costs customers who never return and transactions nobody retries.
Level 1: Personal Finances
The simplest SPOF is one person with one neobank account. The fix is cheap and takes an afternoon.
A minimum redundancy setup
- Two cards from two unrelated providers. A second card from the same bank shares the same backend, so it fails at the same time.
- One traditional bank account. It does not need to be your main account. It needs to work when the app-only bank does not.
- Some cash. Small amounts cover a taxi, a meal or a pharmacy visit while systems recover.
- Offline access to account details. Keep IBAN, sort code or routing numbers somewhere other than the app that might be down.
- A second way to get paid. Freelancers and contractors should give clients an alternate account before they need it.
Common mistake: keeping a backup card in a drawer at home. A backup only helps when it travels with you. Keep it in a separate pocket or bag, so losing one wallet does not take both cards.
Check that the backup works
A card you have never used is unproven. It may be expired, frozen after inactivity, or blocked for online purchases by default. Make a small purchase with it every month or two. If it fails, you find out at a coffee shop, not at an airport.
Stripe makes a similar point for merchants: relying on one payment method is risky, because technical problems or provider policy changes can interrupt business. Alternatives preserve continuity. The same reasoning applies to a personal wallet (Stripe).
Level 2: Payment Flows in a Business
A merchant with one payment service provider (PSP) carries the same risk, multiplied by every customer. When the provider goes down, every checkout fails. When the provider changes its risk policy, accounts can freeze without warning.
How larger merchants handle it
According to the Merchant Risk Council, most larger e-commerce businesses work with 3 to 6 payment providers. They want high availability, acceptable conversion and wide coverage of payment methods. The payments leaders surveyed agreed on one point: avoid dependency on any single PSP if you want to scale.
That does not mean every small shop needs six processors. It means the architecture should allow a second one without a rewrite.
The three parts of a vendor-agnostic setup
Solidgate splits a vendor-agnostic payment architecture into three components:
- Token vault portability. Stored card credentials live in a vault the merchant controls, not in the PSP's vault.
- Routing control. The merchant sets rules by market, card type, BIN range, cost or risk profile.
- A unified API. One integration covers all connected providers.
The first component matters most. If card tokens sit inside one PSP's vault, switching providers means asking every subscriber to re-enter card details. Many will not do it. Without portable tokens, a second processor is only a backup on paper.
Solidgate reports that, per its merchant data, payment cascades across multiple providers recover an average of 14.8% additional LTV compared with single-provider billing. Smart retry logic across providers adds a further 11.6% (Solidgate). These are vendor figures from the vendor's own customers, so treat them as directional. The logic holds regardless: with one provider, a decline is final. With two, it can be retried.
Key point: for subscription businesses, a decline on the only provider means a lost subscriber. A second route turns that loss into a retry.
A simple failover pattern
Orchestration platforms do this as a product. The core logic is short enough to understand, and to build in a basic form:
// Simplified failover across payment providers.
// Each provider exposes the same charge() interface.
const providers = [primaryPsp, secondaryPsp];
async function chargeWithFailover(payment) {
for (const provider of providers) {
if (!provider.isHealthy()) continue;
try {
return await provider.charge(payment, {
idempotencyKey: payment.id, // prevents double charges on retry
timeoutMs: 8000,
});
} catch (err) {
if (isHardDecline(err)) throw err; // card declined: do not retry elsewhere
log.warn({ provider: provider.name, code: err.code }, 'failover');
}
}
throw new Error('All providers unavailable');
}
Three details decide whether this helps or causes harm:
- Idempotency keys. A timeout does not mean the charge failed. Without a key, failover can charge the customer twice.
- Hard vs soft declines. "Insufficient funds" should not be retried on another processor. A gateway timeout should.
- Health checks. Mark a provider unhealthy after repeated errors, so every request does not wait for the timeout first.
BR-DGE adds a caveat that applies here. Any payment system can become a SPOF, including an orchestrator. Their suggested approach is to keep direct connections to PSPs alongside the orchestrator. They call it "belt and braces." If the orchestration layer fails, the direct route still works.
Reconcile, because webhooks lie by omission
Zethic points out that webhooks can fail silently. During an outage, this is the usual failure mode. The PSP processes the charge, the notification never arrives, and the internal database still shows "pending."
The fix is a scheduled reconciliation job. It compares gateway records against internal state:
-- Payments still pending locally after 30 minutes.
-- Feed these IDs to the PSP API to fetch the real status.
SELECT id, provider, provider_payment_id, created_at
FROM payments
WHERE status = 'pending'
AND created_at < now() - interval '30 minutes'
ORDER BY created_at;
Run it every few minutes in production and after every known provider incident. It catches orders that were paid but never fulfilled, and orders marked paid that never settled.
Avoid lock-in at the contract level
Technical redundancy fails if the contract blocks it. Preczn recommends a flexible termination clause without excess fees. It also recommends explicit data portability for customer data. Before signing, ask whether the provider offers tools for exporting data, and whether exports meet your compliance requirements.
Primer notes one more benefit of a processor-agnostic vault: network tokenization through Visa and Mastercard. When a card is reissued or expires, the token can update automatically. That reduces involuntary churn on recurring payments (Primer). Zethic covers the predictable part of the same problem. Notify users 30, 15 and 7 days before a stored card expires.
Level 3: The Infrastructure Behind the App
Customers see a Revolut, a Wise or a neobank logo. Behind it sits a chain of hosting providers, card processors, partner banks and middleware. Any link in that chain can fail.
The Synapse lesson
The Yale Journal of International Affairs analyzes the Synapse collapse as a "weakest-link problem." In Banking-as-a-Service setups, critical services are spread across a chain of fintechs. Some of them are unregulated. One weak point can break the whole system.
Two observations from that analysis matter for anyone holding money in an app:
- Middleware platforms work as invisible infrastructure. They process and hold large customer balances, and consumers are rarely aware of them.
- When failure comes, responsibility blurs. Customers end up stuck between the technology platform and the partner bank, and neither side takes full responsibility.
Cross-border operation makes it worse. A failure in one region can spread to customers in several jurisdictions.
An outage and an insolvency are very different events. A few hours of downtime is not a Synapse. But both expose the same hidden chain, and both are much easier to handle with a second account already open.
What fintech teams should do
A ZyBiSys post on fintech business continuity asks a direct question: is the company ready if its primary hosting provider crashes? Without a business continuity plan, downtime causes lost revenue from stopped transactions. It also brings manual processing costs, regulatory fines and loss of customer trust. Their three key actions are short:
- Redundant hosting. A second region or provider that can take traffic, not just store backups.
- Proactive monitoring. Alerts from outside your own infrastructure, so you learn about an outage before customers report it.
- Frequent testing. A failover that has never been tested is a guess.
Flexential adds the organizational part. Identify SPOFs first, then build a mitigation plan for each. Teams have to be able to disclose weak spots without fear. As Flexential puts it, the goal is not to punish anyone but to build a stable and reliable system. A SPOF nobody admits to stays in the system until it fails.
A quick SPOF audit
For each critical path, such as login, card authorization or payouts, write down every component it touches. Then ask one question about each: "If this disappears for four hours, what happens?"
| Component | Typical SPOF | Mitigation |
|---|---|---|
| Payment processor | One PSP, tokens in its vault | Second PSP, portable vault |
| Notifications | Webhooks only | Scheduled reconciliation |
| Hosting | One region | Tested failover region |
| Status communication | Status page on the same infra | Separate host and channel |
The last row is often missed. If the status page runs on the infrastructure that just failed, customers get silence instead of updates.
Frequently Asked Questions
How do I access my funds if I can't log in to the Revolut app during an outage?
During a full outage, the app is usually the main interface, so options are limited. This is why a second account and card should be set up in advance. Follow Revolut's official channels for updates, and do not use third-party sites offering "workarounds."
Does the outage affect all Revolut users globally or only certain regions?
It depends on which component failed. Some incidents affect a single feature, such as card payments or transfers, or a single region. Official Revolut status updates are the reliable source. Social media reports show where users notice problems, not the scope of the incident.
What should I do if my Revolut card is declined while the service is down?
Switch to a backup card or cash, and avoid repeating the same transaction many times. Repeated attempts can create duplicate pending charges once systems recover. Check your transaction history after service returns and dispute any duplicates.
When will Revolut services be fully restored?
Only Revolut can answer that for a specific incident. Restoration times vary with the cause. Plan as if an outage could last the rest of the day, and be pleased if it ends sooner.
Are customer deposits and data secure during a major outage?
An outage is an availability problem. By itself, it does not mean funds or data are lost. Protection of deposits depends on the entity holding your account and its jurisdiction. Check which Revolut entity serves you and what safeguards apply there.
What to Try Right Now
Try it: take one card other than your main neobank card and buy something small with it today. If it works, put it somewhere separate from your main card. If it fails, you found the SPOF before it found you.
For teams running payments, the equivalent step is just as short. Find out where your card tokens are stored. If the answer is "in our PSP's vault," that is your first redundancy project.
Information is accurate as of the publication date. Terms, prices, and regulations may change — verify with relevant professionals.