We are in an industry where messages are broadcasted to all trading partners (11 of them). From time to time some of those external systems may become unavailable, or are having extensive delays such that they are unable to provide an acknowledgement response in time (delivery mechanism https with timeout of 2 mins). Messages are high volume with about 50K per day.
When those outages occur, most of the available outbound https job deliveries are taken attempting to deliver to the failed party eventually timing out after 2 mins, with substantial backlog for messages to other parties.
My question is: is there a way to build in an externally managed table where trading partners are flagged, such that only those with the UP status would have messages delivered. This would avoid the above problem where we know there is an extensive outage, so that our own backlog can be cleared before turning the flag back on to allow internal TN retry mechanism to take over?
The above idea comes about from having a look at the dbops.sql file, which I believe is internally used by TN to select the relevant jobs for deliveries. I’m hesitant to make changes to this file incorporating the idea above, so just wondering if anyone has safely done this before. Also does anyone have an understanding of the “job polling” frequency from the database?
We are currently on IS 6.1, but this issue has been with us since 4.0.