zevOS

01Platform status

What can break, and what it means for you

Charging platforms have an unusual property: most of their components can fail without stopping a car from charging. Knowing which is which is the difference between an incident and a panic.

Components

Six services, six different consequences

Written so an operations lead can work out, in one read, whether an incident affects revenue or only visibility.

OCPP gateway

The WebSocket service every charger connects to. An incident here means chargers may disconnect and reconnect, and remote commands may be delayed. Chargers with a local authorisation list continue to serve drivers.

Chargers may go offline in the console

Driver charging flow

The QR-to-browser page drivers land on. An incident here prevents new sessions from being started by driver-initiated payment. Sessions already running are unaffected.

New sessions cannot be started by drivers

Payments

Authorisation, capture and refund through the payment aggregators. An incident may be ours or the gateway’s; either way, failed authorisations are recorded rather than silently dropped.

Payments may fail; sessions are not billed twice

Operator console & API

The dashboard and the REST API. An incident affects visibility and remote control, but chargers continue to run sessions and the events are processed when the service recovers.

Reduced visibility; charging continues

Event processing & workers

Session creation, cost calculation, availability updates and alerting. Unacknowledged events are reclaimed on recovery, so an incident here delays processing rather than losing it.

Session and revenue data may lag

Roaming (OCPI)

Inbound and outbound roaming connections. An incident affects partner-initiated sessions and CDR exchange; your own drivers are unaffected.

Roaming partner sessions may fail

During an incident

How we communicate

  1. 01

    We tell you before you ask

    Affected operators are emailed as soon as we have confirmed an incident and know which component it is — not once it is resolved.

  2. 02

    We say what is affected, in your terms

    Whether chargers are still serving drivers, whether payments are collecting, and whether data will backfill.

  3. 03

    We update on a fixed cadence

    Even when the update is "still investigating". Silence during an incident is worse than bad news.

  4. 04

    We write up what happened

    For anything with material impact: the cause, the timeline, and what changes so it does not recur.

Seeing something wrong?

Send the charger identity and roughly when it started. We can usually find the exact frame sequence from that.