__init__(self,
server,
minArchCount=10,
minClientArchCount=1,
cats=None,
maxCount=10000)
(Constructor)
| source code
|
x.__init__(...) initializes x; see x.__class__.__doc__ for
signature
- Parameters:
minArchCount - >=1,default: 10. This is the count of messages that are at
least archived along with any problem message (the problem
message itself included.) Note 1: When a problem is archived
shortly after another, no messages are saved twice even if the
"preceding messages" intervals overlap. Note 2: The
effective number of archived lines is limited by the cached jobs
(job history) of the server. See LogServer parameters.minClientArchCount - >=1, default: 1. This is analogous but addresses messages with
the same client id as the "trigger message". The
intention is that we often want to save more preceding messages
that concern the "problematic" client only. Note:
Unlike with minArchCount, messages are archived twice when the
"preceding messages" intervals overlap, i.e. when
problems are logged shortly after each other.cats - list of cat; I archive log entries when cat is in this
list.Default is "E","I","S" (These
are speaking keys meaning: Error,InternalError,SecurityIssue.
Adapt it to your convention.)maxCount - max. problem count per archiver process. Subsequent problems are
ignored. Note that the size of the archive (and meta table) can
be limited this way, but since the limit is per-process, the max
row count of the meta table can be number-of-processes * maxCount
! (We may consider to make the meta table mandatory; then we
always could query the table size there just to ensure it does
not grow too big. But it seems not worth the effort.) - Overrides:
object.__init__
|