I see. This was very important to know, and it probably points to the resolution of the problem. What might be causing the slowdown of the system might be the fact, that the component internally subscribes (at least for some time) even to the items that you just "read". This help in many scenarios, but not in yours.
In order to prevent this behavior, set the UpdateRates.ReadAutomatic property of your EasyDAClient instance to System.Threading.Timeout.Infinite.

Hi again!

I've trid to use/instantiate a new opcclient localy withing each method and to use
a global static instance used by all threads.

I don't know which variant you recommend?

The problem though I think is as you said with the internal subscriptions to satisfy
my Reads.

>Regards Owe

Hi!

The reads that you do after each "trigger", are they generally to items that you have subscribed to or used before?

The reads after trigger are to items not subscribed to. They will be read for every product entering the station, about every
20 second. We only subscribe to tags that we want to trigger us to do something.

>Regards Owe

Please post information to this forum, unless there is something sensitive that you do not want the world to see. In such case, please post the general part still on the forum, and send the sensitive part separately using some other channel (email).

I would be grateful.
I will send you information tonight. Should I post it on this thread or?
>Regards owe

The reads that you do after each "trigger", are they generally to items that you have subscribed to or used before? If they are not, what might be happening is that more and more items get subscribed internally, for the purpose of satisfying these "reads". If this is the case, I can work with you on tuning the system to better handle this scenario.
The way you have described the characteristics of your application is very good. If there is anything more that can help me understand what the application does and how, please send it as well.

Hi!

Has done some more logging. It works fine in 5-6 minutes. After thar it starts to get
longer responstimes.

The system has approx 3000 tags. We subscribe about 100. We ge triggered some
times per second. After each trigger there is quit a lot reads to execute the usecase.

Could it be that we have to split the system in 2 or more subsystems to hadle the load?

>Regards owe