Skip to main content

Clarification

If a record does not get through and retrying does not help, Xchange holds on to it. It gets a case in clarification — with the cause it is stuck on and with everything needed for the decision. This page answers what happens on its own and how you can tell where you have to step in yourself.

Which values the screen shows in detail is described under Clarification.

Two types, and they mean different things​

Every case carries a type, and it is the first thing you should read:

  • Failure. Something went wrong, and in a way that persists when retried: a mandatory field is empty, a key cannot be matched, the target system rejects the content.
  • Review. Everything ran properly. The job has reached a point at which it is open how to continue, and puts the question to a person.

The type also determines what happens to the record — see the next section.

In the example: WS-2026-104861 is a failure — the order came from the webshop without a customer number, and only through it can the customer be matched. KS-778102 is a review — the status message customs_hold matches no rule of the route that was to dispatch it, so it is open who should handle it. How a message gets there is described under Dispatch.

What an open case does to the record​

An open failure holds its record back. Every further run of the route leaves it aside, instead of letting it fail again and filling the list a second time. The skipped job then shows Skipped (clarification open).

An open review lets it keep running. It only waits for an answer, and the record keeps running along.

A field conflict holds back only the fields it presents. The job writes everything else and leaves only the fields you decide on. In a list this happens per row: the fields of a row are presented one by one, and the other rows go through. If one side has deleted a row that the other has changed, the row is presented as a whole. Which list is compared how is described under Lists.

The block of a failure can end, too. Three things lift it:

  • The record changes in the source system. Then it is a new state instead of the same failure, and the next run takes it along.
  • A deadline expires. If the case has been open long enough, the route takes it along again, although the case stays open. The clarification names the point in time until which it skips it.
  • The case is decided — resolved or closed.

There is one exception: if the job ran into a timeout, only a decision lifts the block. The target system may have started writing, and a second attempt might arrive there a second time. This case waits for a person who can check in the target system.

In the example: WS-2026-104861 stays put, for a different reason: "Bestellungen Webshop" takes its work from a queue, and the message is used up after the first run. Later runs see only new messages — the block has no job to do here.

The cause tells you where to look​

Every case carries a cause, and every cause belongs to a cause group. The group answers exactly one question: In which system do I have to look? It is purely a statement of place, with no statement about who made the mistake or how serious it is.

In the example: WS-2026-104861 is listed under Source data. Nothing is wrong with the mapping of the route; this one record lacks the value for a mandatory field. Looking here means: in the webshop, at the order itself.

Which causes there are and to which group each belongs is listed in Clarification.

Reprocess — and what has to change first​

Reprocessing picks up the record once more, as it stands in the source now, and binds it to the active configuration snapshot. A second attempt therefore ends differently from the first only if the source data or the configuration has changed since the failure.

For a failure, the order is: read the cause first, change something there, then reprocess. What that looks like step by step is described under Resolve a clarification.

For a review, reprocessing is the decision itself: you add what was missing and thereby release the record.

Closing says something about the case​

Closing ends the processing of the case — and the record stays as it is. That is the right way when the job was taken care of elsewhere or is no longer needed at all. It is the wrong one when the record is actually still supposed to arrive: closing only takes it off the list.

How a case ends​

A case is open until it comes to rest in one of three ways:

OutcomeWhat is behind it
ResolvedA reprocessing got through. If a name is shown, someone triggered it; if the name is missing, the record later went through on its own.
ClosedSomeone decided that there is nothing left to do here.
SupersededThere is a newer entry for the same record, for a field conflict already a newer job; the older one is thereby moot.

In between, a case can show Reprocessing: a reprocessing is running right now. If its run ends without resolving the case, it is Open again.

A case resolving itself is the normal case for a route that reads its dataset repeatedly. If the missing detail is added in the source system, the next run takes the record along — and if it gets through, Xchange closes the open case along with it.

Several cases, one cause​

If a configuration fails, it usually affects several records. If there are several open cases with the same cause on one route, the clarification points this out: a change in the right place then resolves all of them at once.

The reverse conclusion holds too. A single case is usually a data problem; many identical cases from a certain point in time on are usually a change to the configuration.

See also​