From time to time we have an app failing due to a connector issue, which causes the set of app triggers to malfunction & fail to either create the entry expected or causes an entry to be incomplete. The process is plain and simple that user gives a signature, which then runs the triggers to obtain data using a connector, update tables and link some records. This malfunction will cause error messages to appear on the user station. As the data entry was incomplete, the subsequent entries over the next days working with the same set of data will cause new new errors to appear. We have had some cases where it takes a few days before IT realize there is an issue and can rectify the issue.
My problem is that we cant monitor if apps on stations have created errors or get notifications if it occur. In station management there are no error logs to inspect so we would rely on end users to open an ticket to IT, which in turn have to reach out to Tulip to help get an explanation.
Apart from instructing users to immediately call upon IT when they see issues, what do you do to handle this?
I am not quite sure if you are referring to internal tulip connector errors, or errors due to the external system server (like http response code 400, etc.), but if is the later (http response code errors), there is this KB article which may be of help -
Another thing, depending on the connector, you usually get a response, and this response should appear in the status or outputs when testing the connector. These can then be added to your connector outputs.
then, in the app, you can read that response from the connector, check if there is an error before proceeding, and then route the user to appropriate steps and notify those who need to know based on that response.
I believe it would help a lot!
If logs are added to a table even automations could help shift through new logging and alert us. It might be that the logs are too complex for people outside the Tulip engineering team and we need help anyway.
I don’t really know how it works for others, but when we detect an issue (because it gets reported to us), we reach out to Tulip support to get help investigating what might have occurred. It means that we cant proactively find the error scenarios. The app designers do their best to make good and stable apps, but there might be the odd case where we would like to investigate immediately and fix issues before next user arrives. Investigations are of cause also intended to help optimize design, and add more error proofing if needed.
We have an open support case from the past months use, where the app triggers didn’t behave as expected & I am merely raising this because we have very little options to debug and catch issues.
