We recommend that you run your production system with some diagnostics switched on. In this way, unexpected problems can be investigated when they first occur. Diagnostics consume resources, and there is always contention between problem determination facilities and performance. You need to decide how important it is to be able to diagnose a problem the first time it occurs rather than having to try to recreate it or having to just turn on diagnostics and wait for it to re-occur.
‘Dump on System Abend’ should always be enabled in an Enterprise Server region, along with some Auxiliary tracing. The following level of diagnostics is the minimum we recommend for running in production systems.
From the Diagnostics tab on the Enterprise Server Admin Console:
We strongly recommend that you also enable the ‘core_on_error’ facility so that application core dumps are created whenever there is a crash: see the topic Producing a Core Dump for Run-time Errors.