Page Comparison

Built with .NET 6.0.10 and Transporter 8.2.0

ConnectorVersion is no more in the configuration, instead MessageBroker reports to FSM Cloud its technical version number

If Transporter fails to start on Engine startup, the startup is re-tried every 30 seconds until successful, if not in failover mode. Behavior in failover mode not changed, if Transporter fails to start, the Engine is stopped and the other instance in FO cluster can take the communication.

SAP web service client state added to GetMessageBrokerInfo. That way it should be possible to monitor from SAP state of all key components

• SAP -> MB connectivity - if not accessible, GetMessageBrokerInfo cannot be called anyway and throws exception

• MB -> FSM Cloud connectivity - Transporter status

• MB -> SAP connectivity - SapServiceClient status

GetMessageBrokerInfo web service operation reports now state of Transporter/Cloud communication

threading optimizations

• Throttling is applied separately to each web service call rather than to each message. Previously throttling could have been applied excessively in case of multi-message (message package) web service calls.

• Throttling is applied to each company separately. Previously it was applied globally (on MB level), which used to cause time-outs on SAP backend side, as the company queues are processed in parallel in SAP.

• Confirmation sending now does not block processing of the normal messages. The confirmations used bo be sent in the main thread, blocking the processing of other messages. Now confirmations are sent in a separate thread.

Messages of type COR_CLOUD_BOUNCE (used by us to mark fullsync-end) are not send to FSM cloud anymore - they are not FSM relevant. They used to be ignored by the cloud, but recently the cloud started throwing errors. Filtering them out should solve the problem.

Confirmation bundling reworked not to lock the communication while the bundle is (synchronously) being sent by Transporter. This should improve performance of the communication in unstable network condition.

First draft of measuring allocated memory implemented. When it proves to be reliable, we may decide to apply throttling based on the memory consumption.

BTP: Added support for location ID of SAP Cloud Connector destination.

Remote MB restart support: stop Message Broker process when requested by the FSM Cloud. OS-level service manager should restart it automatically.

Handling for failover mechanism was added

New message broker makes is possible to remap existing FSM Cloud GUIDs to new SAP keys keeping existing FSM relations in place

Message broker is now capable of presenting it’s own version

In case of configuration errors in MessageBrokerSettings.yaml, Message Broker used to exit without any information, now a meaningful log is produced before the application is stopped

Message Broker 3.2 was built against .NET 6.0

Improved handling of container restart in BTP (required also changes in SAP backend. Check SAP NOTE 3133310

Pinging Message Broker from Connector does not result in ERROR entries in the log. INFO is added instead

In case a proxy is used relevant information can be found in the log