I am writing a system where all events from the manager interface are
logged into a database. This data could be used for a variety of
information gathering:
e.g.
Display of zap/sip/iax channels in use between any period (inbound /
outbound on which channels)
agent details
queue details
messages available in mailboxes
The benefit of this is that I can then let the managers see precisely
what was going on between any points in time i.e.
what was the max load on the isdn30 channels between 2 and 3 last tuesday ?
How many agents were available, and on what queues ?
etc etc
Now, I know that certain information may be available already (such as
cdr, queues etc) but it is a lot easier for me to have a central
repository of the events that took place so that we can re-run the
events as if they were happening all over again.
I have created such a database and system, and it seems to work well for
us. However, it works by capturing the events from the telnet interface.
This is not suitable for our purposes, as the * system may be rebooted
and the connection lost. Once the events are gone, they are gone!
What I would like to have available (don't know if it is or not) is a
mechanism for logging the manager events to a file, just like the other
logs available in logger.c. We would also need to have a unique EventID
logged with each event so that we would know that we have processed that
event already. Again, I know that there are other ways of identifying
the event (at the moment I do a crc of the event "packet" and compare
it
with my database in conjuction with the UniqueID of the channel being
"evented"), but that's very messy.
Does anyone have any thoughts ?
Julian.
