Yossi Dahan [BizTalk]


Thursday, August 07, 2008

Admin console wish list feature

I’ve been fortunate enough to have to go through a bunch of suspended items on our production server from the last week and do some impact analysis – understand what failed, why it failed, what is the impact on the business, and how we can mitigate (technically or non-technically)

I have to admit this is very interesting work, but I suspect only until it gets somewhat repetitive...

Anyway, this has made me come up with a new wish list item for BizTalk, this time for the admin console –

I was going through all the items in the list, for each one I had done some investigation, and possibly raise some queries with others to understand the process/impact/etc.

Getting answers takes a while, so I end up having a bunch of suspended/dehydrated items waiting there until I can get around to resume/terminate them.

To make matters worse - often in BizTalk server you would get more than one instance in the admin console that relates to a failure. the send port might be suspended, the orchestration sending the message might be suspended as a result and another orchestration that called that (or that is waiting for a correlated response) might simply be dehydrated)

All of this made it harder to know, at a glance, where I stand - what have I looked at, what haven't I checked yet, what is this one waiting for and which other process it relates to.

I think there are a bunch of feature that can be added to the admin console to make this work much better.

Things like being able to relate several suspended items together, tagging/colour coding items to categorise them, adding notes to items to indicate where in the investigation process I am etc.

I'm aware that there are several tools out there to help track support tickets etc. but they lack the relation to the tools admin use to monitor the application - in our world this would be primarily the BTS Admin Console!

Labels: ,


  • Hi Yossi,

    One of the biggest limitations I find when doing what you have been up to is actually in HAT.

    It would be really useful to know which server out of your BizTalk group an orchestration shape or message instance executed on.

    Imagine a typical process in a production environment and how it would execute different parts of the process on different servers. You just cant see from HAT what happened where.

    We had a problem once where one of our BizTalk servers in the group had been setup incorrectly and it used to get a problem fairly regularly. When you looked at it at group level it just seemed that the instances of the problem were totally random and intermittent, but once we figured it out it happened fairly regularly on the 1 box that was broken.

    It was such a pain to identify the cause of the problem though.


    By Anonymous Anonymous, at 08/08/2008, 09:05  

Post a comment

<< Home