Moved from GitHub dgraph/3857
Posted by MichelDiz:
What you wanted to do
A massive upsert (milions of nodes).
Why that wasn’t great, with examples
All massive upserts take some time to execute. The massive upsert works fine, but there’s no visual reference during the process if it is going well or not. Sometimes Ratel could fail visually or the Browser (for some reason) and We’re not sure if the Big Upsert was executed or still executing.
We could have two types of reference for this. Could be API responses (if possible) sending percentage values and a log in the Alpha’s log terminal. The log would start to appear only if the transaction takes more time than usual.
This probably would be done after we have a Query Planner.