Could you give me links to "Error handling best practice" arcticles, if any?
The question is when better to use the error handling in orchestarions. At
the first hand, it would be great to have a general error handling
procedure, for instanse, to send email notification to an administrator
and/or write the error to the appropriate log. At the other hand, if the
error has been processed by the handler that means that the orchestration
would not be suspended at the shape which caused the error, and would be
completed successfully, thus, the orchestration messages will be lost.
What could you advise?

[ Post a follow-up to this message ]

I found this article helpful...
http://www.netologi.se/default.aspx...3/22/93981.aspx





"MichaelS" <mmm_shark@yahoo.com> wrote in message
news:eXdkg5GGGHA.344@TK2MSFTNGP11.phx.gbl...
> Could you give me links to "Error handling best practice" arcticles, if
> any?
> The question is when better to use the error handling in orchestarions. At
> the first hand, it would be great to have a general error handling
> procedure, for instanse, to send email notification to an administrator
> and/or write the error to the appropriate log. At the other hand, if the
> error has been processed by the handler that means that the orchestration
> would not be suspended at the shape which caused the error, and would be
> completed successfully, thus, the orchestration messages will be lost.
> What could you advise?
>

[ Post a follow-up to this message ]

I have used the email approach, but what if something like SQL is down for 3
0
minutes, and how process 100 orchestrations per minute?  Are you aread for
3000 emails?

MOM or some other monitoring tool looking at the Event Log is another
possibility.
I think it can be tuned to say something like this - if there are more than
20 exception in a 1 minute period, email me, and/or send me the first
occurence of all messages... etc...

Neal Walters
http://Biztalk-Training.com

[ Post a follow-up to this message ]

I also use the email approach. But rather than directly sending an email
when the error is detected I publish an exception message and then have a
exception handler orchestration which is a aggregating sequential convoy
with a delay shape. This orchestration collects the errors for say 10
minutes then sends an email with alist of the errors. For your scenario this
would mean 3 emails rather than 3000.

Greg


"Neal Walters" <NealWalters@discussions.microsoft.com> wrote in message
news:6D0D8006-2736-42F4-9D1B-3E27EBB22127@microsoft.com...
>I have used the email approach, but what if something like SQL is down for
>30
> minutes, and how process 100 orchestrations per minute?  Are you aread for
> 3000 emails?
>
> MOM or some other monitoring tool looking at the Event Log is another
> possibility.
> I think it can be tuned to say something like this - if there are more
> than
> 20 exception in a 1 minute period, email me, and/or send me the first
> occurence of all messages... etc...
>
> Neal Walters
> http://Biztalk-Training.com
>

[ Post a follow-up to this message ]