Answer guide · English

By Symbioen · Published · Reviewed

What should an EV charging service-level agreement measure?

Keep operator-reported uptime in the service-level agreement (SLA), but do not let it carry the agreement alone. Uptime shows that chargers were switched on. Add measures for whether charging sessions succeeded and how quickly faults were fixed, plus a check on how much of the network produced usable data. For every measure, write down the period, denominator, exclusions and data source, or the two parties will calculate different results from the same clause.

Use case
Procurement, operations and portfolio oversight
Base unit
EVSE (individual charging point)
Independent baseline
Public National Access Point status history
Boundary
Operational evidence, not legal certification

Six measures for a charging SLA

Public EV charger availability
Was the charging point in a working status?
Which statuses count as working, observed time and allowed exclusions
Successful charging sessions
Did charging start and continue?
What counts as a session and whether status, operator or vehicle data is used
Failure and recovery
How often did service fail, and for how long?
When a failure starts, when it ends and the maximum repair time
Signal quality
Is the data timely and stable enough to trust?
How missing, duplicate, delayed and rapidly changing statuses are treated
Utilisation
How much of the observed time was the charger in use?
Context only; low use must not count as a fault
Evidence coverage
How much of the contracted network could be measured?
Charging points with enough data, divided by all charging points in scope

Add Symbioen metrics to the SLA

Operator-reported uptime is a useful input, but operators define it differently and it does not show whether drivers could charge. The measures in the table close that gap.

Most SLAs today are written around uptime. Instead of uptime, we calculate availability: the share of observed time a charging point reported a working status in the public feed. The approach follows the common uptime methodology that ChargeUp Europe proposed in September 2024. We use a different name on purpose, so our figure is never mistaken for the uptime an operator reports from its own systems. If the contract names uptime, say whose uptime, and put availability next to it.

Availability and successful charging sessions describe the service drivers got. Failure and recovery shows how the operator responded when a charger broke. Signal quality and evidence coverage tell you whether the data behind a result really covers the network you contracted for. Utilisation is context: a quiet charger is not a broken one, and the SLA should not penalise low use.

Measure recovery from the moment a fault appears, not from when a technician is sent. ChargerHelp’s 2026 study found about half the recovery time passed before anyone was dispatched. Median recovery: 7 days for a single-visit fix, 32 days when more than one visit was needed.

Define the calculation before you set a target

A small example shows why. Take one charging point over a 30-day month, 720 hours. It reports a fault for 20 hours, is in planned maintenance for 10 hours and sends no data for 6 hours. If the contract leaves maintenance out of the calculation and counts hours without data as working, availability is 97.18%. If maintenance counts as downtime and hours without data are left out, it is 95.80%. Against a 97% target, the same charger passes under one reading and fails under the other.

So agree the definitions first and the target second. The contract should state:

  • the calculation method and its version;
  • which charging points and hours are in scope;
  • which statuses count as working and which as failed;
  • how planned maintenance and missing data are treated;
  • how much data is needed before a result is published;
  • results per EVSE as well as for the whole portfolio (EVSE means an individual charging point).

Targets set by regulators

AFIR, the EU regulation, sets no reliability target. It requires operators to publish data about their chargers but does not say how often those chargers must work. Two other regimes do set a target, and they define it differently.

In the UK, the Public Charge Point Regulations 2023 require each network of rapid chargers (50 kW and above) to be reliable 99% of the time, averaged over the calendar year. A charger reporting an unknown status counts as reliable. In the US, chargers funded by the federal NEVI programme need more than 97% uptime per charging port over the previous 12 months, and utility outages, vandalism and scheduled maintenance can be excluded.

The UK’s 99% and the US 97% are not two levels of the same measure. One averages a whole network and counts missing data as working. The other is set per port and lets the operator exclude some outages. If you borrow either target, borrow its definitions too.

Report each data source separately

Public National Access Point data gives a baseline that works the same way across networks and that neither party controls. The operator’s session and maintenance records show delivered service and repairs more directly. Fleet or vehicle data, shared with consent, shows what drivers experienced.

Report each source on its own before you compare them. A blended figure hides the places where they disagree, and those are usually the ones worth investigating.

What this evidence does not cover

Symbioen provides traceable operational evidence. It can support tender monitoring, SLA reviews and regulatory reporting. It is not legal certification, and it does not confirm that a contract or regulation has been complied with.

Evidence and method

Sources for this guide