Spike in Metarequest call timings in Java client

[Biju N]

Hi There,
    We see spikes in the metadata request calls when a broker machine
specified in the broker list goes down i.e server not reachable. Is there a
property which can be set in the Java client to control the timeout when a
broker is not reachable and try another one in the list. We are seeing this
issue in 10.02 Java client. Any help on this is much appreciated.

Thanks,
Biju