This is one of the most common support cases in charging and one of the most frequently escalated to the wrong party, because both sides have evidence they are fine.
The charger is powered and its indicators look normal. The platform says it has not been seen. Both observations are usually accurate.
Establish which layer is failing
There are four places this breaks: the charger has no network path, it has a path but cannot establish a session, it establishes a session but registration is refused, or it is connected to a different backend than the one being checked.
The last is more common than it should be, particularly after a migration or a service visit where a configuration was restored from an old backup.
Indicators do not report backend connectivity
Most chargers indicate power and readiness rather than platform connectivity. A unit with a healthy-looking display can have been unable to reach its backend for weeks.
This is why the on-site observation and the platform observation are not in conflict. They are reporting different things.
Check in this order
Working out of order produces the familiar loop where the site team and the platform team each demonstrate their side is healthy.
- Confirm the charger is configured with the backend address you expect.
- Confirm it has a network path at all, by signal indication or a local diagnostic.
- Confirm the connection reaches the platform, rather than a captive portal or a blocked path.
- Confirm registration is being accepted rather than left pending or rejected.
- Confirm the identity it presents matches what the platform expects, exactly.
Site network changes are a common cause
A charger on site Wi-Fi stops connecting when the password is rotated, the network is re-segmented, or a firewall rule changes. Nobody associates the charger with the change because the charger is not on anyone's inventory of affected systems.
Where chargers use site infrastructure, they belong on that inventory, and network changes should include them in the impact assessment.
Cellular has its own failure modes
A subscription lapses, a data allowance is exhausted, coverage degrades as a building's occupancy changes, or a network operator retires a technology the modem depends on.
None of these is visible at the charger, and all present as a unit that looks healthy and cannot connect.
Rule out the marginal case
A connection that establishes and drops repeatedly can present as offline when checked and connected in the log, which is confusing for everyone. Look at the pattern over time rather than at the current state.
Marginal connectivity is worth distinguishing from no connectivity, because the remedies differ: one needs an antenna or bearer change, the other needs a configuration or subscription fix.