By default, SapphireIMS credential use timeout as 5 sec with 0 retries. It is advisable to keep the default values. Timeout and reties will be used on each SNMP request and there will be many requests which gets triggered from SapphireIMS to collect all the needed parameters data from target node with different object identifiers (OIDs), so for example, if there are 20 OIDs to be queried on device and all request are getting timed out, then total timeout value will be 20*5 sec = 100 sec and this will double up if retry count is increased by 1.
If random systems are experiencing excessive timeouts and slow in responding, we can create separate credential profile with little higher time out ideally around 10-15 sec MAX and 0 or 1 retries.
If several devices within the network are experiencing excessive timeouts, then probably packets drops (packet loss) might be happening and it is advisable to keep the timeout as default and increase retries.
In any case, it is best to observe the behavior of device as well network and set timeout and retries accordingly.