Decouple ACE Message Flow Monitoring Event publication from the message flow thread

We request an enhancement to IBM App Connect Enterprise (ACE) Message Flow Monitoring to completely decouple monitoring event publication from the message flow execution thread.

Currently, the messageflow, none, and independent transaction settings control how monitoring events participate in the flow transaction, but monitoring event processing remains associated with the message flow thread. Consequently, latency or failures when publishing events to external observability platforms such as Logstash/ELK can reduce flow thread throughput and, under high traffic, impact business transaction latency.

We experienced a critical production incident where Logstash suffered an OOM condition, correlating with severe degradation of ACE business services.

We propose a native asynchronous, fail-open monitoring mechanism based on a separate worker thread or thread pool and a configurable bounded queue. External monitoring failures, network timeouts, or backpressure should never block business processing. Queue capacity, worker concurrency, timeout, retry, and overflow policies should be configurable, with metrics for queue depth, dropped events, failures, and retries.

Although an intermediary such as IBM MQ can provide decoupling, it introduces additional infrastructure and operational complexity.

Observability failure should result in observability degradation, not business service degradation.