Wireless and access points

Monitor → Infrastructure → Wireless describes WiFi from the agents doing the listening. It combines physical access-point groups, their radios, reception across observers, and link health from agents actually associated with them.

This is useful when a controller’s device status does not explain what users experience at their desks. A radio can be powered on and audible while a particular agent’s connected link remains weak.

Collect observations

Configure WiFi Scan checks on agents with a usable wireless adapter in the locations you want to observe. The default interval is 60 seconds. Platform APIs, permissions, adapters, and drivers determine which scan and link fields are available; check the agent’s capabilities and actual test results.

An installed agent without a WiFi Scan check does not populate this view. A scan can hear access points while the agent is wired or unassociated. Those observations provide coverage evidence but no connected-link experience score.

Choose a time window, then inspect Radios and the reporting observers. Search by SSID, BSSID, or access-point name. Signal includes the reception spread across observers: the same radio may be strong near one agent and weak near another.

Group radios into physical access points

A BSSID identifies a radio/network advertisement; one physical access point can broadcast several BSSIDs across bands and SSIDs. Grouping connects those radios to the unit you name, inspect, and place on a floor plan.

There are three ways to create or extend a group:

1. Auto-group associated groups eligible radios associated with your agents. Unassociated neighbouring radios are left for review.
2. Suggested access points proposes groups from shared base-MAC patterns. Review the radios before accepting a suggestion; a naming pattern is evidence, not independent proof of hardware identity.
3. Select ungrouped rows in Radios to create an access point or add the radios to an existing one.

Expand an access-point row to inspect its radios. Edit the group to change its name or membership. One grouped access-point pin on a floor covers all its member radios. Deleting a group returns its radios to the ungrouped list and removes that group’s floor-plan pins.

Separate experience from coverage

Experience uses the link an agent is actually associated with. Coverage uses the radios it hears. A strong neighbouring access point can produce good coverage while the agent is using a weak link to a different radio.

Health summaries include contributing agents and samples, good/fair/poor counts, and unassociated scans separately. Depending on platform evidence, they can include mean and worst-decile RSSI, SNR, transmit rate, and retries.

The current grading is reproducible:

• RSSI of at least −67 dBm is good; below −75 dBm is poor; the range between is fair.
• When SNR is available, at least 25 dB is good and below 15 dB is poor. A worse SNR grade can lower a sample that looks acceptable from RSSI alone.
• The score is the percentage of graded samples that were good, with fair samples counting half. No graded samples means no score, rather than zero.

Unassociated scans are shown separately because they do not describe a connected link. Retries and transmit rate add context but are not the grading formula. A negotiated transmit rate is not measured application throughput, and a high score does not prove end-to-end application health.

Investigate a wireless complaint

1. Select the relevant window and find the access point or radio serving the affected agent. Verify its actual association where reported.
2. Compare experience with coverage. A loud radio nearby may be a neighbour or a radio the agent did not choose.
3. Inspect the reception spread and observer count. One strong observation does not describe the whole floor.
4. Compare the mean with worst-decile signal and available retries. Intermittent trouble can be obscured by a broad average.
5. Use floor plans to inspect where samples were pinned or estimated, then compare the agent’s service and path tests.

Radio channels and bands help investigate overlap, but a scan alone is not a complete spectrum-utilization measurement. Bluetooth LE Scan adds beacon observations for positioning on supported Linux and Windows agents. Advertiser counts can provide context about the radio environment; they do not directly measure WiFi airtime or prove interference.

When the view has no health score

First check whether scans exist. If they do, distinguish “heard this radio” from “associated with this radio.” Wired agents with scanning adapters can provide coverage while supplying no experience grade for an access point.

If expected radios are absent, clear grouping/search filters, widen the window, and check the WiFi Scan result and capability report. Review grouping membership before interpreting a physical access point’s aggregate. Continue with Infrastructure troubleshooting.