We have SMD installed primarily to do proactive
monitoring and regular health checks for the system.
What I want to open up here is a discussion on how we
monitor the Portal systems using SMD?
I asked that question a week ago also...
How can we classify the errors whether they are
critical or can be skipped?
No way to do that - Portal monitoring is "reactive", never proactive.
Generally it is that we face a issue and solve it
<b>not</b> pro actively monitor. I tried my best in
getting materials from the Web for the same but was
helpless. It is terrifying to note that with many
MNCs using portal as production servers we don't have
a pro active monitoring methodology other than get a
issue and face a down time.
That's bane and bless of the new technology. ABAP needed 10 years to become what it's today, J2EE is compared to that a fetus, at most an infant; I'm sure SAP will do best, to make it as robust as ABAP but as of now, running a J2EE engine productively needs (quote from an OSS message) "... substantial knowldege of J2EE internals...".
And given that fact (amongst others), we are really doing brainstorming about our strategy and whether it's not negligent to invest time and money in a technology, that is from a customer point of view almost uncontrollable and reactive only. We actually use the portal as a a nice OPTIONAL frontend for certain information but I would have a real problem, if we'd use it as the only single entry point to get valuable company, financial and statistical information.
I believe there are lot of System administrators
active in forum request you all to kindly contribute
your views as to how we can pro actively monitor
portal server.
yes - absolutely.
One more question : After how many days should a
portal server be restarted?
The best thing is to not restart it at all - as you usually do with your ABAP systems. But given the fact, that the Service Marketplace (frontend?) webservers are also restarted frequently in the OSS (I often get "Service not available" during the day) throws an interesting light because I vaguely remember, that they run (a modified) EP too...
I have posted that kind of criticism often and through various channels, it might appear, that I want to just talk mad about SAPs J2EE, but that's not the case. I just see the daily challenges for SysAdmins because I'm sitting "at the front", not at the green table somewhere working nine-to-five doing nice powerpoint presentations. I'm sure,things would be different, if some of the developers would be at the customers site and experience a LIFE SysAdmins working day with all the things, that arise, including monitoring and troubleshooting of various systems, technologies and issues.
And for me the most paradox thing is, that "Business Process Monitoring", which is in all mouths and brains, means monitoring something across system boundaries (which can also include an Enterprise Portal). As of now I was not able to talk to ANYBODY who is capable, of telling me, what components of a portal must be monitored to make sure, a certain business process is running. Does application A needs e. g. SLD, is do I need to make sure UDDI/JMX/RMI/Jwhatever is up and running and doing nicely and IF it´s part of that business process and how can it be monitored?
--
Markus
Message was edited by:
Markus Döhr
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.