Treat your hosting exit plan as a business responsibility

Your website can be running correctly when you lose the ability to operate it. A provider may suspend an account, a commercial dispute may escalate, or the only administrator may lose access. A backup inside that same account is little help if you cannot retrieve it.

For a small business, the sensible response is to preserve a way out before buying a second copy of every service. Keep domain control and recoverable backups outside the hosting account, understand where DNS depends on the same provider, and rehearse recovery. Separate DNS where the added operating work is justified. The goal is a business that can move without the departing provider's cooperation.

The May story was an allegation, not a verdict

On 26 May 2024, Robin Dev published an account of a dispute with Cloudflare. The author identified himself as an online casino's systems engineer. He alleged pressure to buy an annual Enterprise contract for $120,000 and said the company's domains were removed on 16 May, affecting its website, email and internal access systems. The post's publication date and the alleged removal date are different.

That is the customer's account. It does not independently establish Cloudflare's motive or contractual fault. This article does not resolve those questions. Looking back in November 2026, the useful business question is how many essential functions could become unavailable together if one account were withdrawn. That question applies regardless of which party was right in this dispute.

Count dependencies by what you would lose

Ask the person responsible for your software to trace a visit to your website. Domain registration preserves your right to use the name. DNS directs traffic to services. Hosting runs the site, while a content delivery network, or CDN, can sit in front of it. Email has its own provider but depends on the domain's DNS records.

Then trace recovery. Which email address receives account reset messages? Where are deployment instructions stored? Does the administrator need a login service controlled through the same account? A list of suppliers can look diverse while the recovery route still passes through one locked door.

The most useful inventory records who can change each dependency, where the recovery credentials are kept and what remains accessible after suspension. An agency's assurance that it handles everything does not answer those questions. The business should know who holds authority over its domain and who can exercise that authority if its usual technical contact is unavailable.

Separate the controls that let you leave

Start with domain registration at a registrar whose account is independent of hosting and whose rules allow the nameserver changes you need. Protect that account with strong authentication and recovery details that do not depend solely on the affected domain's email. Restrict routine access while retaining an authorized recovery route.

A different registrar is not automatically enough. Cloudflare's current Registrar documentation states that registered domains must use Cloudflare's authoritative DNS. This is a current constraint checked for this retrospective, not a claim about the service's exact rules in May 2024. It illustrates why you must inspect the control you retain, rather than count the products on a bill.

DNS can remain with the CDN provider if the business knowingly accepts that dependency and can rebuild it elsewhere. Separating authoritative DNS creates another operating responsibility; it should not be treated as a free improvement. For a small company, independent registration, an off-provider DNS export and a tested move may be the appropriate first step. The remaining delay should be explicit.

Keep application data and recovery material outside the account whose failure you are planning for. Include the files needed to rebuild, deployment instructions and a controlled way to recover required secrets. Check that the copy can actually be restored. The existence of a paid backup feature is evidence of a feature, not of a successful recovery.

A DNS export is only part of a move

Cloudflare documents how to export DNS records. That provides a useful record to keep elsewhere. A zone export is not evidence that your application, access policies or traffic rules can run at another provider. Those need their own recovery instructions and replacements.

Moving traffic also changes the protection around the application. Do not assume that exposing the original server is a safe emergency shortcut. A particularly concrete dependency is the certificate: Cloudflare warns that Origin CA certificates can produce browser trust errors when its proxy is bypassed. That warning comes from current documentation. A recovery plan should establish a valid route to the replacement service before relying on a DNS change.

A technical owner should also account for DNS caching and DNSSEC, the mechanism that authenticates DNS responses. Changing nameservers without coordinating its trust records can break resolution, as Cloudflare's migration guidance explains. A lower record lifetime does not guarantee that every visitor changes route immediately. The business needs a tested estimate with assumptions, rather than a universal promise of recovery within a fixed number of hours.

Rehearse the failure you are buying protection against

A useful rehearsal assumes the original provider's dashboard is unavailable. Using only the separately held material, ask the technical owner to restore a recent backup into an isolated environment. Check a representative customer task and establish how traffic and email would be redirected. A test hostname can cover much of this work without moving the public site.

Record what the rehearsal proves and what it leaves untested. Restoring the application does not prove that public DNS delegation has changed successfully. A copy that is a day old also does not recover orders accepted after that copy. Decide how much data loss is tolerable alongside the time needed to resume service.

Put a name beside the person authorized to make the move, the person doing the technical work and the person confirming that customers can use the result. Include a way to communicate if company email is affected. These are business decisions that give the technical instructions a chance to work under pressure.

One provider can still be the right choice

An integrated platform can reduce maintenance and let a small team deliver more useful work. Running several providers increases billing, access management and configuration work. For a brochure site whose owner can tolerate an interruption, a recent export and a documented rebuild may be enough. A service taking orders continuously may justify more preparation.

The recommendation is proportional control. Use the platform that serves the business, but budget for the dependencies it creates. Include a recovery rehearsal in software maintenance and repeat it when the architecture changes materially. A credible exit plan ends with evidence of what can be recovered, by whom and under which conditions.