Orangehat Business Solutions

IT Infrastructure

Cloudflare's new beaconing signal puts context ahead of incident labels

OrangeHat Editorial·2026-11-17·3 min read

A connection that repeats on a schedule can give an IT team a useful question before it supplies an answer. Cloudflare's new periodic-beaconing signal makes that distinction explicit, offering a timely example of how a detection should enter an investigation without becoming an unsupported incident label.

A navy laptop connects to an orange server block along a path of teal discs, with a magnifying glass over the path and a blank card nearby.

What Cloudflare announced

Cloudflare's November 16 changelog introduces Periodic Beaconing Traffic as a Cloudflare One risk behavior for Enterprise plans. It examines Gateway HTTP traffic for regularly spaced contacts to the same destination, using traffic behavior rather than threat intelligence, categories or destination reputation.

Cloudflare describes a regular pattern that few colleagues share and cautions that legitimate software can produce it.

The behavior is off by default with an Info risk level. Events usually arrive three to six hours after the six-hour analysis window ends, according to Cloudflare.

Begin with a bounded question about the traffic

For a business IT team, the practical question is what produced the observed pattern and whether it matches an understood activity. The detection's label alone cannot resolve that question.

A hypothetical first review could record the relevant time window, destination, affected device and the application expected to communicate with it. An authorized administrator could then compare those observations with approved software and maintenance activity, keeping sensitive details inside the organization's normal investigation process.

This example describes a review approach rather than an observed customer incident. A plausible explanation would still need supporting evidence; an unfamiliar destination would likewise need investigation before being described as hostile.

Account for when the signal becomes visible

Because the announced behavior uses an analysis window and delayed reporting, the event's arrival time and the time of the underlying traffic can differ. A handoff that preserves both makes it easier to ask the right operational question.

For example, a hypothetical investigator receiving an event after a maintenance period would want to compare the event's underlying window with that period. Looking only at what is running when the event appears could leave the earlier activity unexplained.

The once-per-continuing-beacon behavior also limits what an event count can establish. A count of notifications is a different observation from a count of network contacts, and it should be labeled accordingly.

Detection settings and response decisions need separate ownership

Cloudflare's risk-score documentation describes controls for enabling behaviors, changing their assigned levels and using user risk scores in Access policies. Those are configuration choices with operational consequences, rather than evidence that a particular event is malicious.

A team evaluating this release could define who reviews the signal, what corroboration is required and how an authorized response decision is documented. Any policy change belongs in the organization's established change process.

The immediate contribution is a more focused lead for investigation. Its value depends on whether the team can connect that lead to relevant evidence and explain the resulting decision without overstating what a traffic pattern proves.

What to watch

Keep the observation and the conclusion separate

A useful investigation record explains the repeated traffic, the evidence used to interpret it and the reasoning behind any response. The new signal can help frame that work; it does not establish that an organization is compromised or that a particular destination is malicious.