Cloudflare can make a WordPress site faster and safer for free. It can also cause redirect loops, broken logins, and a cached shopping cart showing someone else’s items — if it’s set up carelessly. The difference comes down to a handful of settings.
This guide shows how to run WordPress with Cloudflare the right way: why the pairing works, how to connect your domain, which SSL settings to use, caching rules for WordPress, and security features worth enabling.
New to CDNs? Start with What Is a CDN and Does Your Site Need One?.
www) to Proxied (orange cloud). Keep mail-related records DNS only (grey cloud).Weighing Cloudflare against your registrar’s DNS? See Cloudflare DNS vs Registrar DNS: Which Should You Use?.
This is where most problems happen. Under SSL/TLS → Overview:
Make sure both URLs in WordPress Settings → General use https://. Need a certificate on your host first? Read What Is SSL? How to Get HTTPS for Free.
By default, Cloudflare caches static files but not HTML pages. If you add full-page caching, you must bypass the cache for:
/wp-admin/* and /wp-login.phpwordpress_logged_in)Easier options: Cloudflare’s APO for WordPress (verify current availability, settings, and the actual application behavior) or a caching plugin with Cloudflare integration. Purge the Cloudflare cache after major design changes. For plugin options, see Best WordPress Caching Plugins Compared.
Also install the official Cloudflare plugin or configure your server so WordPress sees real visitor IPs instead of Cloudflare’s. Many hosts do this automatically; otherwise security plugins and comment logs may show only Cloudflare addresses.
Record the current DNS zone, origin certificate, redirect behavior, and cache ownership. Imported records require review: a scan is not proof that every mail or verification record was copied. Coordinate DNSSEC delegation when changing authoritative providers; an old DS record cannot simply be assumed valid for a new signer. Follow the official DNSSEC key guidance for the selected migration route.
For origin TLS, verify that the certificate and hostname meet the Full (strict) requirements. Do not resolve certificate errors by weakening validation as a default workaround. Investigate the certificate chain, expiry, hostname, and origin connection instead.
Cloudflare does not cache HTML or JSON by default; see the documented cache behavior. If full-page caching is introduced, test separate anonymous sessions and personalized routes. Confirm that account, cart, checkout, and authenticated content remain private under the exact rules used. A plugin setting alone does not establish that every custom endpoint is covered.
Promotional recommendation: Consider KAWEDMA website development and hosting services for help reviewing a scoped WordPress integration. Confirm DNS, certificate, caching, testing, and recovery responsibilities in the quotation.
Affiliate disclosure: Hostreta may earn a commission if you purchase through this link. hosting.com managed WordPress hosting is an infrastructure option to investigate. Verify how any existing edge service interacts with a separate Cloudflare account; do not assume the configurations can be combined without provider guidance.
Draw the path from the browser through DNS, the edge service, and the origin application before changing settings. Identify whether the host already provides an edge integration, reverse proxy, or managed cache. Two services with overlapping responsibilities can create confusing behavior if nobody knows which layer owns redirects, certificates, or invalidation.
Record the actual domain names, public routes, mail provider, and administrator responsible for each component. Keep secrets and account credentials outside the general change log. WordPress with Cloudflare should be an understandable architecture rather than a collection of dashboard toggles copied from several unrelated tutorials.
Choose representative checks: a public article, an image, a login, an authorized form submission, and any store journey that matters. Define the expected result before making a change. A working homepage is useful evidence, but it does not prove that every application route or background integration remains functional.
Compare the records imported into the destination zone with the existing authoritative configuration and the application’s requirements. Include mail exchange, sender authentication, verification, service aliases, and less frequently used subdomains. Do not assume an automated scan discovers every required record, particularly an entry that is not publicly queried during the scan.
Document which hostnames should be proxied and which require a direct DNS route. Follow the actual service provider’s instructions for mail and other protocols. If a hostname serves several purposes, resolve the architecture before simply switching its proxy status. WordPress with Cloudflare still depends on correct records for systems that do not pass through the web proxy.
Plan the change with the domain owner and keep a recoverable record of the previous configuration. Coordinate signing and registrar delegation using the documented DNSSEC migration procedure. This article describes planning checks; it does not authorize changing a live domain or weakening its validation without an appropriate scoped plan.
Check the origin certificate, supported hostname, expiry, and connection before relying on strict validation. Identify which layer issues HTTP-to-HTTPS and hostname redirects. Several individually sensible redirect settings can still form a loop when their assumptions differ, so test the complete public path rather than reviewing each setting in isolation.
When a certificate error appears, record the exact public hostname, time, and error. Investigate the origin configuration and certificate requirements instead of treating a weaker encryption mode as the normal repair. WordPress with Cloudflare should preserve the intended validation boundary while the actual defect is corrected.
Keep direct-origin testing separate from ordinary public testing and use only an authorized, supported method. A certificate intended for an edge-to-origin relationship may have a different trust context from a browser-trusted public certificate. Do not publish private origin addresses or credentials in a support example merely to illustrate the diagnosis.
List the layers involved: browser cache, edge cache, server cache, application cache, and any plugin-managed behavior. Decide which layer handles each job and how it is purged. A change that appears missing may be served from a layer that the editor’s purge button does not control.
For personalized content, test more than one isolated session through the supported workflow. Review account screens, cart and checkout, query-dependent pages, and custom endpoints. Do not infer privacy from one successful request. WordPress with Cloudflare needs rules that match the actual application, including extensions and custom routes that a generic recipe may omit.
Record the visible response behavior and relevant cache indicators where available. A cache hit for one image does not prove that the page HTML is cached, and a cache miss is not automatically a defect. Define what should be cached and why before treating every uncached response as a performance problem.
Inventory payment callbacks, scheduled requests, monitoring, editor access, and third-party services that contact the site. Review the supported authentication and network requirements of each integration. A security challenge that is useful for a human login page can interfere with an automated service that cannot complete that interaction.
Test changes with representative authorized traffic and keep an escalation route if an important integration stops working. Avoid broad exceptions that remove protection from unrelated routes. WordPress with Cloudflare requires balancing the application’s legitimate communication with the intended control, rather than disabling security whenever a single request fails.
Where visitor addresses are restored at the origin, follow the provider’s trusted-proxy guidance. Do not trust a client-supplied header from every network source indiscriminately. Keep the implementation with the responsible administrator and document how the origin distinguishes trusted proxy traffic from other requests.
Open the actual public routes in the browsers and devices relevant to the audience. Check images, navigation, forms, error states, and any interactive feature that depends on scripts or APIs. A staging preview may not use the same domain, edge rules, cookies, or certificates as production.
Compare performance tests under stated conditions, including the connection, device, cache state, and page version. Do not present a single improved laboratory score as proof of an equivalent visitor or revenue improvement. WordPress with Cloudflare can support delivery work, but the actual effect must be evaluated on the site being maintained.
Use our Core Web Vitals guide and mobile testing checklist for complementary checks. Keep technical observations separate from assumptions about search ranking. A CDN configuration alone does not establish relevance, indexability, or a guaranteed traffic increase.
Record what changed, which account owns the configuration, the previous values needed for recovery, and the checks performed. Avoid relying on a screenshot alone when the team needs a clear explanation of the intended behavior. Keep recovery material protected and accessible to the authorized people who may need it.
Define what would trigger a rollback and which dependencies must be considered first. DNS, certificates, cache rules, and application URLs do not always revert safely as one undifferentiated action. WordPress with Cloudflare should have a change plan that explains the specific route back, including any delay or verification needed after a reversal.
Review the arrangement after a host migration, domain change, commerce extension update, or major theme change. The initial configuration may remain active while the application evolves around it. Assign an owner and review date so the edge integration does not become an undocumented dependency.
Record the entry URL and the observed destination sequence through an authorized diagnostic method. Identify which layers issue redirects and which scheme each layer believes it receives. Check origin TLS and the application’s supported proxy configuration. Avoid changing several unrelated settings at once, because that makes the successful correction difficult to identify and reproduce.
Confirm that the change was saved and that the public page uses the expected template and assets. Then identify the cache layer serving the older version. Purge through the supported route and verify the result. Do not assume that every stale stylesheet requires a DNS change or that disabling all caching is the appropriate permanent solution.
Check the integration’s supported logs and the relevant request outcome without exposing secret tokens or customer data. Compare the failure with recent rule changes. Determine whether the request reached the application and whether the application accepted it. A browser-visible form and a server-to-server callback can follow different verification paths.
Imagine a consultancy with a public service site and a contact form. This is a planning example, not a measured Hostreta customer result. The team inventories the current records, confirms the origin certificate, identifies cache ownership, and chooses a representative public page and authorized form test. It records the expected behavior before the change.
After implementation, it verifies public delivery and confirms the form reaches the responsible person. If one image is served from an edge cache, the team records that specific observation rather than claiming every request is accelerated. It keeps the previous configuration and a clear recovery owner for follow-up.
Keep the WordPress with Cloudflare answers beside the dated configuration record. A new maintainer should be able to understand the purpose of the integration without guessing from several active plugins. The objective is a tested, recoverable delivery path, not the largest possible number of enabled features.

No. Choose it only when the application, privacy boundaries, and maintenance process support it. A site can use DNS or static-asset delivery without introducing the same HTML caching behavior. Review the workload and the supported integration rather than assuming that every available feature belongs in the setup.
Ask the host for the supported architecture before changing delegation or proxy settings. A managed integration may own part of the delivery configuration already. Document the boundary between the services and test the actual result. Do not assume that stacking services automatically improves performance or security.
No. A cache hit or miss for one request describes that observation. Privacy verification requires representative routes, session states, and the exact application behavior. Keep tests authorized and avoid using real customer data merely to demonstrate a configuration. Record the cases checked and the limitations that remain.
Provide the supported method for publishing changes, the relevant purge workflow, and the contact responsible for delivery issues. The editor does not need unrestricted access to every infrastructure setting to update an article. Keep operational instructions separate from secret credentials and explain when a problem should be escalated.
We may earn a commission on purchases made through these links, at no extra cost to you.
Don’t miss the chance to participate in the biggest giveaway of this year.