Handle faulty indices, by synchronizing timeouts

Currently, the timeout for a cluster to respond is the default of 30s. However, on the server side state changes are allowed up to 120s to propagate, see publishing the cluster state. By increasing the response timeout we avoid unnecessary app failures by preventing eager timeouts.

Bug: T356933 Change-Id: I2f6722f44dee252024b124079d496c4aee4d8b0c