What an undelivered decision means
When the operator approves or rejects a request, the system passes the decision on to the portal. If the portal is temporarily unavailable or the connection fails, the decision does not arrive. The request has been settled in the plant, but the contractor does not yet see the outcome in the portal.
The system then saves the decision with the request, together with an error description. Nothing is lost. Approval of a request is atomic: the firm, the account, the authorisations and the pass are created in one transaction, and an error at any step leaves the request unresolved. An undelivered decision therefore concerns only the communication with the portal, not the consistency of data in the plant.
How to tell a network error from a refusal
This is the basic skill of the duty officer. A network error or an unavailable portal concerns the delivery of the decision, while a refusal is a substantive decision of the operator. The error description saved with the request lets you tell them apart: the first contains a technical cause, the second is just the status of the request.
To be certain, check the number of errors from the last 7 days in the integration panel. If there are many, the problem probably lies with the connection or the configuration, not with a single request.
- The error description on the request states the technical cause.
- The request status says what the operator decided.
- The number of errors from the last 7 days shows the scale of the problem.
Sample shift review (minutes)
Resending from the panel step by step
Resending a decision is limited to the service administrator role. The duty officer opens the request with the undelivered decision, reads the error description, makes sure the cause has been removed or has passed, and starts the resend. After success the error description disappears, and the contractor sees the decision in the portal.
Good practice: do not retry in a loop. If the first attempt failed for configuration reasons, retrying will not help and will clutter the error report.
Shift routine and monitoring
Add a check of resends to the routine of every shift: at the start, open the integration panel, look at the number of errors from the last 7 days and the requests waiting for a decision, and resend the undelivered decisions. The whole review takes a few minutes.
In the monthly protocol, record the number of resends and their causes. If the same cause keeps coming back, report it to IT as a corrective task, instead of retrying every time.
Idempotency and safety of resends
Resending is safe because the system behaves predictably: resending the same request identifier returns the same result, and the same identifier with different data is rejected as a conflict. No duplicate accounts or passes are created.
In the event of a conflict, do not force the resend. Check whether the person's data have been changed, and consult the administrator. It is a signal that someone is trying to use the same identifier for different data.
Communication with the contractor
A contractor waiting for a decision does not see the technical causes. We recommend a short message from the plant: the request has been processed, and the answer will appear in the portal once the connection is restored. Such information reduces the number of calls to reception.
If a pass is urgently needed, the operator may pass the information to the contractor through another channel in line with the plant's procedure. Entry to the site remains the plant's decision, whether or not the portal has already received it.
Checklist
Video
Frequently asked questions
Does an undelivered decision mean the request is lost?
No. The decision is saved with the request, together with the error description, and can be resent from the panel.
Who can resend a decision?
The service administrator role, the same role that approves and rejects requests.
Can a resend create a duplicate?
No. Resending the same request identifier returns the same result.
Let's talk about the contractor portal at your plant
We will show you on a live example how a contractor's request reaches the operator and how the plant keeps full control over entry to the site.
Let's talk