Website Maintenance & Hosting
WordPress Under a DDoS Attack? Respond Without Blocking Customers
A suspected DDoS needs traffic and origin evidence, targeted containment, and business verification. Distinguish hostile floods from legitimate spikes or slow backend work, then test forms, logins and payments before removing temporary restrictions.

In this guide
If WordPress is unavailable during a suspected DDoS attack, first ask the host and CDN to compare traffic with origin resource failures. Apply the narrowest effective containment at the layer receiving the attack, then verify that customers can still browse, submit forms and complete payments. High traffic or a slow website alone does not prove a DDoS.
A distributed denial-of-service attack aims to exhaust availability. It can overwhelm network capacity or target expensive application requests. The response depends on that distinction: a WordPress plugin cannot independently solve an upstream network flood that prevents requests reaching the server.
Confirm the incident from more than one signal
Record the first failure time, affected pages, customer errors and recent changes. Check whether the problem coincides with a promotion, import, backup or release. Those can cause pressure without hostile traffic.
Ask the technical team to compare a shared time interval across CDN security events, origin access logs, application errors and hosting resource metrics. Useful observations include the request paths and methods, request rate, response codes, CPU or memory pressure, database contention and exhausted worker capacity.
- Traffic rises and expensive endpoints saturate: inspect request patterns and application-layer mitigation.
- The host reports upstream bandwidth or connection pressure: escalate network-layer handling to the host or protection provider.
- Traffic stays normal while backend work stalls: investigate slow queries, application failures or background jobs.
- The site fails only for some visitors: compare security blocks, geography, networks and cached errors.
Our high-TTFB diagnosis guide covers the backend-performance branch. A DDoS response should not conceal a failing application that remains slow after the traffic subsides.
Why short incidents still need retained evidence
Cloudflare’s Q3 2026 threat report describes a shift toward shorter attacks. Across its own network, it reports mitigating 18.2 million network-layer attacks and 14.4 trillion HTTP DDoS requests during that quarter. Those are provider-wide observations, not WordPress-specific statistics or evidence about your site.
A burst may end before someone opens the dashboard. Preserve time-bounded events rather than deciding there was no incident because the site now loads. Keep the timezone and distinguish request counts from visitor counts; they measure different things.
The owner can document customer impact and coordinate staff. The host or developer should preserve technical evidence and identify the constrained layer. Assign one incident lead so several people do not make conflicting firewall and cache changes.
Contain the pattern while preserving essential access
Use the CDN or upstream provider’s supported protections where appropriate. Confirm that the affected hostname is actually routed through that service. A dashboard rule cannot protect traffic that reaches an exposed origin through a different path.
The developer should identify the smallest useful scope: a particular abused endpoint, request pattern or action. Consider rate limits or challenges with the business’s normal traffic in mind. Record the rule, reason, start time and a condition for changing it. Avoid treating an entire customer country or network as malicious without evidence.
Cloudflare’s DDoS protection FAQ explains aspects of its mitigation systems. Product capabilities and available controls depend on the service in use. The host should identify what protection applies to your actual network and application routes.
Our WordPress WAF rollout guide covers careful policy changes. During an incident, the same discipline matters: retain a record and test each significant change against essential customer journeys.
Use emergency challenge modes deliberately
Cloudflare describes Under Attack mode as a last-resort measure for layer 7 attacks. It presents a browser interstitial and can affect access and analytics. It can be applied selectively; enabling it across the whole site is a business-impacting decision.
Browser challenges need special attention around non-browser callbacks, APIs, uptime checks and payment notifications. A payment processor cannot complete a human browser challenge as a shopper would. Have the developer preserve required trusted integrations through supported narrowly scoped configuration, without creating broad security bypasses.
Do not disable all protection because one legitimate request failed. Locate the responsible rule or layer and correct that exception. Also confirm visitor IP attribution if application-level rate limits suddenly treat many customers as one visitor.
Verify business continuity, not just a green uptime indicator
- Open representative public pages from independent networks.
- Test an ordinary enquiry form and confirm receipt.
- Verify authorized staff login and essential administration.
- For a shop, run the agreed safe checkout test and match the order with its gateway result.
- Check payment callbacks and connected integrations for delayed or blocked delivery.
- Confirm origin errors and resource pressure improve alongside the traffic mitigation.
For store-specific testing, use our WooCommerce test-order checklist. Monitor genuine user failures as well as blocked traffic. A higher block count is not automatically evidence of a better customer outcome.
Remove temporary restrictions with a plan
After stabilization, review whether an emergency rule is still needed. Relax temporary restrictions in controlled steps while watching origin health and legitimate requests. Reconcile delayed enquiries, notifications or orders, and retain the evidence supporting the incident assessment.
A useful handover states the affected period, likely attack type, observed limits, active controls and monitoring owner. It should also identify whether exposed origin access or expensive application behavior needs a permanent change.
FAQs
Does a sudden traffic spike prove a DDoS?
No. A campaign, crawler, broken client or expensive backend task can produce similar symptoms. Correlate request behavior with security events and origin pressure.
Can a WordPress security plugin stop a network flood?
It operates after traffic reaches the application environment. Network saturation requires help from the host or an upstream protection service.
Should I leave Under Attack mode on permanently?
Evaluate its customer and integration impact. The vendor describes it as an emergency measure; use a documented exit plan and more targeted ongoing controls where suitable.
Does a DDoS mean the database was stolen?
An availability attack alone does not establish data theft. Investigate separate evidence of unauthorized access or changes if it appears.
Sources
- Cloudflare: Q3 2026 DDoS threat report
- Cloudflare: DDoS protection FAQ
- Cloudflare: Under Attack mode and limitations
Prepared for website owners by Nithu / Suthakaran Nithursan. Dated advisory facts link to primary sources. Examples are illustrative and are not client results or guarantees.