Is it possible to direct flow from Write to DSE Smart Service on error?
Ideal scenario would have the DSE node fail, and based on error, flow would continue to different node.
Right now process pauses when DSE fails and email goes out. Is it possible to continue process flow on DSE fail?
Discussion posts and replies are publicly visible
No. The DSE node does not support this. With 22.3 the new Write Records node allows you do disable the default error behaviour and allows you to design a different path in case of errors.
May I ask what you would want to do in case of an error? I almost never wanted to deviate from the default in the past.
In case of error (specifically a duplicate entry error), we would like to end the process. We can not check before the DSE node for duplicates because the node itself is the race condition. It is possible multiple processes would be started with duplicate data, in which case, first process to reach DSE would pass, all others would fail with duplicate errors.
Got it. If this is a new application I suggest to go with the records approach. Else a stored procedure to insert the data might be an option. It could catch the error and return an error state for you to know whether it could insert the data or not.
I believe there is a Feature Request somewhere for allowing to choose the "Pause on Error" setting true/false on Write to DSE, as we have in the older Query DB node Peter Lewis?
My biggest use case for this, is when you write to a database that may be offline on occasion, expected or unexpected. In this scenario (with the Query DB node), I can route them all to a Receive Message node on error. Say 100 process instances error, and move to the message node. Once the DB is back online, I can trigger a message to all of these processes waiting with 1 click and they all will re-attempt the DB write and continue on successfully.
Our planned feature for 22.3 that Stefan mentioned above is basically our approach to meeting this use case. Our intention is to add capabilities to the Write Records smart service to continue to make writing to the database easier for a variety of scenarios, and that's why we added the pauseOnError option to that smart service.
If you haven't seen how it works yet, I'd recommend looking at the recent product announcement webinar or stay tuned for the 22.3 release in a couple of weeks!
In write to data store entity smart service (in process model), only output is "stored values", I want to retry 3 times if the write to data store entity fails, how can I determine failure in this case?
The basic idea is that if a write fails, only a sys admin can analyze and resolve the root cause, and then trigger the node a second time.
This has changed a bit with write records, as it allows you to handle a failed write.