Could Cloudflare rule changes make bots hit my site harder?

November 20, 2026 4 answers
Asked by Nexluno.4j

After changing some Cloudflare bot protections, traffic jumped from roughly 8,000 to 10,000 daily pageviews to 35,000, and origin CPU stayed near 100%. A lot of requests appear to target invalid pagination URLs that end up at the homepage.

Could bots be reacting to the rule changes, or is there a better way to figure out what caused the spike?

Answers

Ruvmexo-3c

The timing alone doesn't establish that the bots noticed your changes. Some scrapers retry with different IP addresses or browser profiles when blocked, but a traffic surge can also be unrelated or expose a caching problem that was already there.

Compare Cloudflare analytics and origin logs before and after the changes. Look at request paths, response codes, cache status, firewall actions and request rates. Then check which requests actually reached the origin and line those up with the CPU spike. Pageview counts alone won't tell you how much work WordPress did.

Reply to this answer
5r!gasketspool_739

Those invalid pagination URLs are a good place to start. If each unique URL misses the cache and WordPress handles the redirect or renders the homepage, a pile of junk requests can keep PHP busy even while the site looks fine to visitors.

Check the actual response and cache status for a sample of those paths. Where practical, return a real 404 for invalid pagination and cache it at the edge, or handle an appropriate redirect there. Test the rule carefully so legitimate pages and logged-in traffic aren't cached or redirected by mistake.

Nexluno.4j:

That gives me something to check beyond just trying to block every bot. I'll look at how those URLs are handled and whether they reach the origin.

Reply to this answer
Mivzaro.9t

For the login alerts, hiding the fields behind a URL parameter isn't much of a barrier. A bot can still post directly to the login endpoint, and the parameter could show up in logs or referrers.

Consider removing that script and using an edge challenge or rate limit on login requests instead. Review traffic to XML-RPC too, and block it if the site doesn't need it. Be careful with broad REST API limits because legitimate site features may use those endpoints.

Reply to this answer
6k+parcelspanner_395

I'd treat Under Attack mode as a temporary pressure valve, not proof of what caused the traffic. Compare the requests being challenged with what still gets through, then adjust one rule at a time so you can see which change affects origin load. Challenges can also inconvenience legitimate visitors, so keep an eye on that.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions