A charger can appear online on a dashboard while a resident still cannot start a session, authenticate, pay, reach support, or get a useful status update. Define uptime as the resident service the building intends to deliver, then agree how each component will be measured and what happens when it fails. There is no universal uptime target or automatic migration right for every apartment network.
Define the service and its measurement
Before comparing service-level terms, decide what a resident must be able to do. This may include starting an authorized session, receiving a clear status, obtaining a session record, contacting support, and using an agreed fallback. Ask the supplier to define:
- which stations, ports, hours, and user types are included;
- how availability is measured and how often station status is reported;
- whether failed authorization, payment, communications, or a blocked connector count as unavailable;
- which planned maintenance, building power events, and force-majeure events are excluded;
- when outage response begins, who can dispatch a technician, and how residents are informed;
- what remedies, service credits, or escalation apply if agreed measures are missed.
Set the target from the building’s use case and contract, not a percentage copied from a different public network or fleet program. A residential service used every night has a different consequence from a visitor charger that is used occasionally. For example, the U.S. Department of Energy’s Alternative Fuels Data Center summarizes California uptime and reporting standards for qualifying chargers installed on or after 1 January 2024 and bought through a state incentive program or ratepayer charges: a 97% uptime requirement and at least six years of uptime-data reporting. Details vary by charger type and power, the number of chargers at a site, and site ownership; the standards do not apply to residential dwellings with fewer than five units. Check the current California rule and funding agreement before applying any of those terms to a building; this is not a default target for apartment networks.
Keep different service outcomes separate. Track port availability, successful authorization, charge start, charge end, completed session, payment and receipt, and help or recovery as different measures. For each one, define the time period, covered stations and users, numerator, denominator, exclusions, and which failed attempts the system can observe. ChargeX customer-focused KPI guidance proposes measures such as start, end, and session success; its OCPP mapping also notes that a backend may not see every attempt, depending on how a driver starts. Use these as prompts for a measurement plan, not as a residential SLA or a target copied from public charging.
One California multifamily report counted 85 chargers at analysis start and 102 at end, with 124 installed by completion; it lacked a full year from all 124. Compare report IDs and dates with the cohort; site count alone does not prove full coverage.
Test what happens when communications fail
“Offline” can mean an EVSE cannot reach the internet, a shared modem or router has failed, the carrier is unavailable, the management service is down, the payment provider is unreachable, or the building controller cannot communicate with the chargers. These conditions can have different effects and may share one dependency. Name the owner and recovery contact for each component, and test the full path from station to resident-facing payment and support.
Require a demonstration for the exact equipment and software. Ask what remains local: recognizing an approved access token, applying safe preconfigured charging limits, recording a session, showing a fault, and restoring queued data after reconnection. Ask what stops, which actions require a live service, and how a resident requests help. Do not promise that charging will continue until the vendor has demonstrated that behavior and the electrical designer has confirmed safe controls.
Test local credentials as separate cases: add, remove, and revoke a resident credential while connected, disconnect the station, try each expected offline behavior, then reconnect and check synchronization and the session record. As one manufacturer-specific example, Bender’s current implementation documentation says its controller uses a fixed authorization list; it may accept SendLocalList without changing that list, while offline authorization remains available. That behavior is not an OCPP-wide rule, so verify the exact model, firmware, and local configuration offered for the building.
Great Britain’s smart-charge rules provide one jurisdiction-specific example: covered private charge points sold for domestic or workplace use in England, Scotland, and Wales have device-level requirements that include continued charging when the communications network is unavailable. The guidance excludes Northern Ireland, public charge points, and some other product cases. It is not a global rule for every building network. Public-charge regulations are a separate regime and depend on how a station is intended to be accessed.
Prepare the supplier change before there is a problem
An OCPP interface does not automatically move a station to a new operator. A supplier change can require new configuration, accounts, network addresses, certificates, SIM arrangements, firmware, and tests for each equipment model. The Open Charge Alliance’s migration guidance describes transition examples and the need to minimize disruption; certification may help with migration but does not replace a building-specific contract or compatibility test.
Ask the outgoing and incoming providers to agree a transition package containing:
- station ownership, model, serial number, port, location, firmware, and current configuration;
- administrator accounts, credentials, certificates, ownership transfer, and the method for changing the management endpoint;
- session and meter history, data format, export rights, API limits, retention, and deletion dates;
- tariff, access, resident account, payment, and support responsibilities during the change;
- warranty status, replacement parts, update access, and any recurring fee that survives termination;
- a test plan, rollback contact, cutover window, outage notice, and post-change acceptance record.
Test one representative station of each model and firmware with the incoming system before moving the whole building. Confirm that the new service can recognize users, enforce required controls, report session data, and restore support. For a critical resident service, put transition assistance and downtime remedies in the contract while both providers are still available.
For the handoff, compare sample exports with the receiving system by stable station, port, session, and resident-account identifiers. Reconcile missing, late, duplicate, corrected, and out-of-order records against the source period before cutover; a continuing data feed alone does not show that the new operator received a complete, usable history.
Check who the rules apply to
A station restricted to residents and their visitors may be treated differently from one intended for the general public. U.K. guidance lists multi-occupancy residential premises with shared parking and resident-or-visitor-only areas among examples that are not public charge points. It also says access during restricted hours alone does not make an otherwise public station non-public. These examples are specific to that regulatory context; check the current rule and actual site access in each jurisdiction.
Use the companion guides for protocol and pairing tests, system requirements, and charging-data protection. If measured energy appears on a resident bill, also check the relevant U.S. meter-verification guide or Canadian approval guide. The building’s daily maintenance and resident operating policy should also say who handles access and service issues.
Checked: 9 November 2026. Network support terms, regulations, and available product features can change; verify the contract, site classification, and exact equipment before relying on a fallback or exit plan.