This was by design. Please find the reason that our Product team has provided below.
Snaps decide which error are written to error views and which errors cause pipeline failure. The reason the decision was taken initially to fail pipeline on endpoint connectivity errors was so that we do not repeatedly retry an operation when the endpoint is down. So only data errors go to error view, connectivity errors cause pipeline failure. So if you have a million documents going to a DBInsert, we do not retry database connect a million times if the database is down
We currently have user stories open on this issue. Will keep the thread posted. Thank you.
