If my application is responsible for redirecting/doing a single sign on to a destination managed by a third party, in general, where should I draw the line for error handling during this process? If an error happens on the other application’s end, is it reasonable for my stakeholder to expect the application I am working with to share responsibility for handling these scenarios?
Notes:
- I am going to keep solutions limited to those that entail only one
request–I am familiar with the “do an xmlhttprequest and see how
that fares before doing anything else” approach. - I am speaking in
terms of an enterprise-level application with fairly decent customer
traffic.
1
In my experience, customers are only happy when the entire solution works. Not that it partially works because of some 3rd-party bug. A fair portion of my working life is working around bugs – in other systems, in the code libraries and frameworks I employ, inconsistent data, the client’s own systems.
You can stop working around bugs when the customer is happy with the solution I think. 🙂
1
You are not responsible for the error yourself, but most customers would hold you responsible for trying to fix/tpatch the error.
-
The first thing you should try to do is to notify the other party of the error at their end.
-
The second thing you might try to do is to to fix the error from your end (to the extent possible).
-
The minimum you should do when redirecting traffic to the other site is to point out the error at the other end to users. Pylons and flashing lights do this from time to time when there is, say a hole in the road.
-
If all else fails, you should consider “pulling” your directions, so customers aren’t directed to the erroneous site through you.