Thank you for the info about tag count etc.
And, good that the Node Id part is resolved.

The "subscriptions stopped" problem is different from the Node Id or such issues - it is an internal OPC UA communication problem that (at least in theory) should never be caused by anything that you do in your app.

In OPC-UA terms, one "subscription" holds and handles multiple monitored items. If the subscription parameters of your monitored items are all the same, QuickOPC will create just one subscription. If this is the case, the subscription Id would be irrelevant to you. If the parameters of the monitored items differ (such as their sampling rates), there will be multiple subscriptions, but I do not think this is the case.

Can you please try to slow down all sampling rates that you use significantly - to 1000 milliseconds or so? I understand that this is not what you want from your project, but I am just trying to determine whether the issue would go away in this scenario.

Best regards

How many monitored items are we talking about, and how often in reality they change?

The subscriptions are real. The client creates a subscription in the server, then adds monitored items to it. The server only sends relevant (changed) data to the client. On the (OPC UA) service level, a "long polling" approach is used, which in the end effect is equivalent to no polling, it is a true subscription: The client issues a Publish request to the server, and the server does not return until it has a new data (or other condition, such as related to keep-alive, occurs). This approach is chosen in OPC UA so that it can be used with transports like HTTP which do not allow "callbacks".

Regarding the error with 'BadNodeIdUnknown': I think that this is either unrelated to the "subscription stopped" problem, or it (less likely) can be somehow related on the server side, as a consequence of "subscription stopped". It is very unlikely that we would "make up" (generate) this error on our side, OR that we would "spoil" the Node Id being passed to the server, causing it to legitimately return this error. I therefore think that either 1) the particular Node Id is truly invalid (given the configuration of the server), or 2) this is a server problem, where the server rejects a Node Id that is valid and should not be rejected.

To the "The OPC-UA subscription published has stopped." error in the monitor change event: This is the reflection of the entries you have seen in the log file, therefore essentially the same primary issue. Thank you - I just wanted to verify what and how gets reported to your app, because *that error* is the actually the description of the problem you have, and the log entries you posted initially then just serve as an explanation for that problem.

How many monitored items are we talking about, and how often in reality they change? I wonder whether we are not dealing with a load that is over the capabilities of the server, the client, or the connection in between.

Thank you