most executions take less than 10 seconds to run but approx 10% take up to 60 seconds to start.
the process allows simultaneous executions, the load is relatively low, and low latency mode does not resolve the problem.
the problem has occurred on both nodes.
seems to affect all listener processes
the issue has only observed in an environment which is running a molecule.
the molecule has two well sized nodes and an elastic load balancer
memory and disk usage statistics haven't indicated any issues
has anyone seen this kind of behaviour before?