Here you have the ability to define a data group for the timer. For
each data record in this data group, a corresponding timer event will
be generated. The number of data records can be restricted with a
filter.
Process each record in a separate transaction has the effect
that one access will occur for each individual data record. This
function is recommended when processing very large amounts of stored
data, such as for a data replication via process, as errors will be
skipped and the transaction duration (and with it, the duration of
database locks) will be kept short.
This setting can also be used to send mass e-mails. With it, one can set
that in the case of any errors that occur, the entire sending process
will not fail, but rather that mails will only fail to send in the case
that an error occurs.