Spark starts crashing when some of the eventhub partitions go down.

[pablo-statsig]

We have seen three times in the past few months where one or more of the partitions in our eventhub go down due to some internal Azure problem. The rest of the partitions keep working, and data is written to the working partitions. However, during this time spark starts to crash with a Futures Timed out error. Is there a way to have spark continue to work even when it cannot connect to some of the partitions? Is there a manual way to set some partitions not to be read from during these outages?