A public EV charger that looks fine but will not deliver a charge is worse than useless, because a driver has driven to it counting on it working. Uptime is the reliability measure that captures whether chargers are actually available when drivers arrive, and uptime monitoring is the practice of watching a network continuously to know which stations are working, which are faulted, and which have quietly gone offline. This page defines uptime as a reliability KPI for public charging, including the federal NEVI expectation, explains what counts as downtime, and describes how a monitoring system detects stuck or faulted chargers, tracks session success, and triggers remote resets.
EV charger uptime monitoring in one line: EV charger uptime monitoring is the continuous tracking of whether the chargers in a network are available and able to deliver a charge, expressed as an uptime or availability percentage over time. Downtime includes faulted connectors, stations that have gone offline, and chargers that accept a driver but fail to complete a session. Federal NEVI-funded chargers are required to meet a 97 percent uptime target, so monitoring detects faulted and unreachable chargers, measures the share of sessions that succeed, and can trigger remote resets to recover a stuck station before a technician is dispatched.
Uptime is simply the fraction of time a charger is available and capable of charging a vehicle, and it has become the headline reliability metric for public charging because it reflects the thing a driver cares about most: will this charger work when I get there. A network can have plenty of chargers on a map, but if a meaningful share of them are down at any given moment, the driver experience is unreliable and the network's reputation suffers. Measuring uptime turns that experience into a number an operator can manage, set targets against, and be held to.
For chargers built with United States federal funding under the National Electric Vehicle Infrastructure program, uptime is not just a good practice but a requirement. The NEVI rules set a minimum uptime of 97 percent, calculated as an average over a year, which means a funded charger is expected to be available the overwhelming majority of the time. This has pushed the whole industry toward taking uptime measurement seriously, because meeting a defined percentage requires actually knowing, minute by minute, whether each charger was up or down, and being able to demonstrate it. The requirement makes uptime monitoring a compliance function, not only an operational one.
Because uptime is an average over time, brief outages that are quickly recovered hurt far less than long ones that go undetected. A charger that faults and is reset within minutes barely dents its yearly uptime, while one that quietly goes offline and stays down for days can single-handedly pull a site below target. This arithmetic is exactly why fast detection and fast recovery matter so much: the goal of uptime monitoring is not only to record downtime accurately but to shorten it, catching a problem the moment it happens rather than when a frustrated driver reports it.
Downtime is broader than a charger being switched off, and defining it correctly is the heart of honest uptime measurement. The most obvious form is a faulted connector, where the charger has entered an error state and refuses to start a session, whether from a ground fault, an overtemperature, a communication error with the vehicle, or an internal hardware problem. A faulted connector is down for a driver even though the station may have power and its screen may be lit, so uptime accounting has to treat the faulted state as unavailable rather than counting the charger as up just because it is powered.
A second form is a station that has gone offline, meaning it has lost its connection to the management backend. An offline charger cannot be authorized, monitored, or controlled, and from the operator's point of view its true state is unknown. Depending on how uptime is defined, an unreachable charger is often counted as down, both because it cannot be confirmed working and because a driver who needs remote authorization or a network app may be unable to start a session on it. Losing visibility of a charger is therefore a reliability event in its own right, not merely a monitoring inconvenience.
The subtlest form of downtime is the failed session: a charger that appears available, accepts a driver's plug-in and authorization, but then does not deliver a charge, whether because it errors out at the start, stops partway, or never begins ramping power. A station can score well on simple availability while still failing a significant share of the sessions attempted on it, so serious uptime monitoring also tracks session success rate, the proportion of started sessions that actually deliver energy and end cleanly. Counting only whether a charger reports itself available can mask exactly the experience drivers complain about, which is why the strongest definitions of reliability combine availability with whether sessions truly succeed.
The raw material for uptime monitoring is the telemetry the chargers already report, and in an OCPP network that means the status notifications, fault messages, and transaction events flowing to the backend. By watching status notifications a monitoring system knows the moment a connector enters a faulted state, and by watching the connection itself it knows when a charger stops communicating and has effectively gone offline. Combining these gives a live map of which chargers are available, which are faulted, and which are unreachable, updated as events arrive rather than on a slow poll, so a problem surfaces in near real time.
Session success is tracked by following each transaction from start to finish. When a session starts, the system watches whether meter values show energy actually being delivered and whether the session ends normally rather than aborting with an error. Over many sessions this yields a success rate per charger and per site, and it exposes a station that is technically available but repeatedly failing to deliver a charge. That signal is often more actionable than availability alone, because a charger that faults every third session needs attention even though it never shows as fully down, and only session-level tracking reveals it.
Detecting a problem is most valuable when it leads to a fast fix, and a common first step is the remote reset. Many charger faults are transient and clear with a reboot, so a monitoring system that sees a stuck or faulted charger can trigger a remote reset through the management protocol and check whether the charger returns to service, all without dispatching a technician. If the reset does not recover the station, the same monitoring escalates to a work order with the fault details already captured. A cloud SCADA and site-monitoring platform such as Merobix can sit on the OCPP stream to do exactly this, presenting live availability and session-success across a network, alerting on faulted or offline chargers, triggering or supporting remote resets, and keeping the historical uptime record that both operations and requirements like NEVI's 97 percent target depend on. The chargers report the facts, and the platform turns them into detection, recovery, and the documented uptime an operator has to prove.
Chargers funded under the United States National Electric Vehicle Infrastructure program are required to maintain at least 97 percent uptime, calculated as an average over a year. Meeting that target requires actually knowing, over time, whether each charger was available or down, and being able to demonstrate it, which makes uptime monitoring a compliance function as well as an operational one. Because it is a yearly average, brief outages that are quickly recovered hurt far less than long undetected ones.
Downtime includes faulted connectors that refuse to start a session, stations that have gone offline and lost their connection to the backend, and failed sessions where a charger accepts a driver but does not deliver a charge. A charger can be powered with a lit screen and still be down for a driver, so uptime accounting treats the faulted and unreachable states as unavailable. The strongest definitions also factor in session success, since a station can appear available while failing a large share of the sessions attempted on it.
Many charger faults are transient and clear with a reboot, so a monitoring system that detects a stuck or faulted charger can trigger a remote reset through the management protocol and then check whether the charger returns to service, without sending anyone to the site. If the reset recovers the station, downtime is minimized; if it does not, the same system escalates to a technician work order with the fault details already captured. This detect-and-reset loop is a major reason continuous monitoring keeps uptime high.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.