We have an environement where 3 CICS regions run on 1 Lpar. Basically they pass the transactions off from one to the next and then back out to the user.

In one of the regions, as volumes increase the CICS response time increased. DB2 response times went down and WLM shows no CPU dispatching issues and there are no storage/enqueue issues.

We tracked this down to an increase in CICS Redispatch wait time for a particular transactions. Most of ours gave a redisptahc wait of 0 to .05 as does this transaction. However as the volume picks up this transaction sees the redispatch waits rise to over .2 seconds.

First, does redispatch wait mean the amount of time waiting to restart running under CICS after coming back (from DB2 in this case)?

Secondly, what can be causes of this wait time in a environment with no CPU/Storage shortages?