IF there is a message and queue, and the run detector dies, and another instance is started, check that the messages are consumed straight away.
Currently it has been observed that if the activemq client dies, it is still shown as connected for multiple hours, and messages are not recieved by the new client.
When switched to a topic/subscription model this was not the case.
According to the ActiveMQ documentation, the point to point model should never allow messages to be lost, but a loss of messages is possible with the topic subscription model.
We must identify the exact issue, as we do not want to lose messages in production and have to implement a way to find missed runs, post fact.
IF there is a message and queue, and the run detector dies, and another instance is started, check that the messages are consumed straight away.
Currently it has been observed that if the activemq client dies, it is still shown as connected for multiple hours, and messages are not recieved by the new client.
When switched to a topic/subscription model this was not the case.
According to the ActiveMQ documentation, the point to point model should never allow messages to be lost, but a loss of messages is possible with the topic subscription model.
We must identify the exact issue, as we do not want to lose messages in production and have to implement a way to find missed runs, post fact.