Messages show in Pending queue quite often

We have been working with Software AG folks for 10 days now on a critical production issue. Let me put this out in the forum and see if anyone else have encountered the same issue.
We are on IS9.12 clustered via Terracotta stripe with UM on 9.12(single instance) (latest build). What we see time to time is that IS publishes the data to Universal Messaging. No build up of CSQ. On and off throughout the day, the Pending column displayed by Enterprise Manager shows message count as big as Max Pending and then no messages gets picked up. So the receiving trigger does not fire. The problem with all this is UM is up, IS is up. no warnings or error.
Software AG has been unable to make any progress on this and our corrective action is to edit the Trigger in question and do a edit/save and then it clears it out of Pending and rolls over to other clustered IS instance and processes from there.
This is clearly a bug. Has anyone else seen this errant behavior?

Its been frustrating to say the least without any solution in sight and we are really regretting upgrading from Broker to UM.
Any insights or help someone can provide would be fantastic.

Regards,
Nathan.

Are you sure that you are using a single instance of UM (no A-A or A-P) servers in PROD?

We are using IS 9.12 with two IS clusters (A-P Terracotta servers) and 3 nodes of UM servers. No issues reported currently as we are on latest fix levels of IS and UM.

You can connect me via email (in signature) if you want to have a discussion on this.

Thanks. What I listed as 2 IS and 1 UM is indeed what we have.We are not planning to add UM node arrays at this time. I don;t believe this is a requirement from what i can see.

Hi,

Have you reached out to PD expert “Jonathan Heywood” and get his expert view as it sounded very critical issue?

HTH,
RMG