A Transaction is a vehicle used in the flow of Messages on a bonded ticket. It is essentially a container for a Request/Receipt pair. The Transaction record contains the necessary fields to facilitate and track both the movement and the processing of the pair of Messages.
A Transaction record’s primary functions are:
To contain the Request/Receipt pair of Messages.
Handling the queueing of Requests and Receipts.
Tracking the state of the Transaction from two perspectives:
- The condition of the Transaction from a movement (transactional) perspective (i.e. the condition of the transport of the data).
- The condition of the Transaction from a process perspective (i.e. its relation to the business logic that’s in place).
System A sends a Request to system B to update the value of the
Urgencyfield on a ticket without having ‘ownership’ of that ticket. The Request is properly formed and authenticated and is received in system B.
System B sends a response acknowledging the Request and then subsequently processes the Request asynchronously. If, on processing the Request from system A, system B discovers that system A did not have the authority to change the value of the
Urgencyfield, they would send a Receipt Message stating the error. That Receipt is also properly formed and authenticated and is received in system A without any problems.
Transaction statein this case would be set to ‘Complete’ as both the Request and the Receipt were transported successfully. However, the
Process statewould be set to ‘Rejected’ as system A did not have the authority to update the value of the
One of the many defining features of Unifi is its ability to track the state of the Transaction from both a transport (What’s happening with the movement of the message?) and a process (How does the content of the message align with business logic?) perspective.
Having the combination of these two states is extremely useful and gives a far more precise understanding of the condition of the Transaction and opens up more options. This means, for example, that an analyst can see and has the opportunity to fix issues based on a process error without the need for technical support (Unifi won’t break an Integration for a process error). For transactional errors Unifi can throw events which can be correlated to send notifications, or trigger the system’s Incident process as appropriate.
|Sending||0||An outbound Message is being sent to the integrated system.|
|Received||1||An inbound Message has been received from the integrated system (but not processed).|
|Awaiting Receipt||2||An outbound Message has been sent to the integrated system and we are awaiting an asynchronous Receipt.|
|Sending Receipt||3||We are sending an asynchronous receipt in response to receiving and processing an inbound Message from the integrated system.|
|Received Receipt||4||An inbound asynchronous Receipt has been received from the integrated system.|
|Complete||5||The Message exchange was completed.|
|Queued||6||The Message exchange has been temporarily delayed because of incomplete transactions.|
|Timed Out*||7||An outbound Message has been sent to the integrated system, but we have not received the asynchronous Receipt in the expected timeframe.|
|Error||8||There was a transactional error with the Message exchange (e.g. invalid message format, unknown endpoint, code error etc.).|
|Ignored**||9||Transactions can be ignored manually to allow the integration to proceed.|
This is configured on the integration by the value in the
Async timeout field.
This can be used when a Transaction does not complete (i.e. when it is in an ‘Error’ or ‘Timed Out’ state), or at any time when a Transaction is in progress or ‘Queued’ (e.g. when a system error occurs outside our control).
|Pending||0||The Message is being processed and is awaiting a decision as to whether it falls within or outside of the scope of the business logic.|
|Accepted||1||The Message has been processed and accepted as within the scope of the business logic.|
|Rejected||2||The Message has been processed and rejected as outside the scope of the business logic.|
The following table gives a description of the fields that are visible on the Transaction record:
|Message ID*||String||The unique internal message identifier.|
|External message ID*||String||The external system’s unique message identifier.|
|Table||Table name||The target table this record applies to.|
|Document||Reference||The integrated ServiceNow ticket.|
|Integration||Reference||The integration this record belongs to.|
|Bond||Reference||The bond this record belongs to.|
|Message||Reference||The message used to process this record.|
|Direction**||Choice||The direction this record is travelling in.|
|Transaction state||Choice||The state of communication for this transaction.|
|Process state||Choice||The business logic process state.|
|Created||Glide date time||The date and time the transaction was created.|
|Receipt due by||Glide date time||The date a receipt must be received before this transaction times out. Only applies to asynchronous transactions.|
|Error||String||The internal communication error.|
|Process error***||String||The business logic error in processing this transaction.|
These are used in Unifi to match the related Requests and Receipts (i.e. the Message ID of one system will be used as the External message ID of the other system and vice versa).
The direction of the initiating (or first) Message in the Request/Receipt pair will determine the direction of the Transaction.
The value in this field can be rolled up to the bonded ticket, giving the analyst insight and the opportunity to resolve process errors.
Attach logs field in the Integration is checked, then the transaction logs will be attached to the record and you will be able to view the logs from within the Transaction record itself. This hugely significant feature offers considerably more detail than is available with the OOTB ServiceNow system logs and puts that information right where you need it - saving an inordinate amount of time hunting through the
These outbound logs are from when the ticket was updated, through to creating the Transaction and Stage.
These outbound logs are from taking the Stage data, building the Request record & sending the Request to the integrated system.
These logs are from processing the inbound Message.
Transaction next queued
Once the initial Transaction has completed, these are logs from checking whether there are any other transactions queued and processing those.
Replay (UI Action)
This will be used when testing/debugging. For example, if there is a transactional error, after you have investigated the cause & made any required changes you will want to replay the Transaction to check if the error clears.