Technical operations6 min read

OCPP Transaction Troubleshooting: From Accepted Command to Session Evidence

Mayank Harsani
OCPP Transaction Troubleshooting: From Accepted Command to Session Evidence

Troubleshoot OCPP transaction issues by separating command acceptance, transaction records and energy evidence. Follow a safe, version-aware investigation.

A driver says charging did not start, while an operator sees an accepted request. Those observations can both be true. A useful investigation follows the evidence across the customer action, backend, charger and final record. The situations below are illustrative troubleshooting scenarios, not Sock8 incident reports, measured failure rates or instructions to change a live charger.

An accepted request is one piece of evidence

The EV Roaming Foundation’s OCPI examples include a remote-start request that is accepted but produces no session because a vehicle is not connected before timeout. The operational lesson is to avoid using a command acknowledgement as the only proof of a successful customer charge. Collect the downstream transaction and device evidence for the same attempt.

Identify the protocol before reading the sequence

OCA explains that OCPP 1.x reports transaction information across StartTransaction, StopTransaction and related metering/status messages. OCPP 2.0.1 uses TransactionEvent for transaction-related reporting and introduces configurable start/end points. Its transaction IDs originate at the charging station, unlike the CSMS-assigned IDs described for 1.6. Use the matching specification and errata; do not apply a 1.6 log checklist unchanged to 2.0.1.

Build one timeline before changing configuration

  1. Identify the attemptRecord a redacted customer reference, site, charger, connector or EVSE, model, firmware, protocol version and a narrow time window.

  2. Align the clocksRecord device timestamps and backend receipt times with time zones. Preserve original order and note gaps rather than silently correcting them.

  3. Correlate the recordsMatch request identifiers and transaction/session references across the relevant systems. Keep retry attempts distinct.

  4. Preserve the evidenceCollect the applicable request/response pairs, state changes and metering records. Redact personal tokens, credentials and payment data before sharing.

Scenario: request accepted, no transaction observed

In this illustrative case, first check whether the missing transaction is genuinely absent or merely hidden by the interface time range or identifier mapping. Review the command result and subsequent device reports, then verify vehicle/connector readiness and the relevant authorization result with the supplier’s documented behaviour. Check connection interruptions and delayed delivery before deciding that the device failed. Escalate with the timeline and exact configuration rather than repeatedly issuing a start command.

Scenario: a transaction exists, but energy or price looks wrong

Compare the meter samples, their units, timestamps, scope and relationship to the transaction. Distinguish a running register from an interval value using the applicable protocol fields and device documentation; do not subtract arbitrary dashboard numbers. Then compare the metering interpretation with the agreed tariff and billing rules. A billing discrepancy can originate after the protocol exchange, so include the calculation and payment records without treating them as interchangeable with raw charger evidence.

Scenario: the session view remains open after interruption

Check the device’s reconnect history, retained records and the final transaction information received by the backend. Determine whether there is a missing message, a late message, a duplicate or a display/reconciliation issue. Follow the documented recovery procedure for that charger and software version. A reset or forced record closure can change the evidence and customer outcome; involve the responsible operator before taking action.

What a message validator can and cannot establish

A syntax or structural check can catch malformed input within the tool’s documented scope. It cannot show that the charger accepted a message, that the full transaction sequence was valid, that electricity flowed or that billing is correct. Use a redacted example for the check and retain the original logs securely. Verify the tool’s supported version/actions and do not describe a lightweight check as OCA certification.

A better escalation package

Send the narrow timeline, redacted protocol messages, exact device profile, expected versus observed behaviour and reproducible steps. Name any recent configuration or firmware change and who authorized it. State what remains unknown. Keep physical and electrical diagnosis with the appropriately qualified people; a backend log alone cannot diagnose every site fault.

Ready to build your charging network?

Tell us where you are today. We'll recommend the appropriate platform and deployment model.