Proper way of handling terminal stream failures with JMS connector + IBM mq? #3160
Description
With jms connection factory configured to use client auto reconnect + auto acknowledge + manual acknowledgement strategy under MQ, each reconnection will count as a failure inside a multi inside JMSSource, and after the failure count exceeds the value specified in retry.max-retries the consumer will stop being "ressurected". There seems to be no clear way of detecting that this happened other than polling the HealthCenter class, then it will give a general "not ok" sign and it will be possible to figure out which specific consumer was killed. Is It possible to restart a specific consumer indefinitely or am I supposed to count on the application's health check endpoints to give a 503 so that the host (i.e kubernetes) can pick It up and restart It? I can also terminate the application inside a multi that checks the liveness status every few seconds and have no issue with doing that, but I want to know If this is the way to handle terminal failures (which are really just benign reconnection events)